Internet of Things (IoT): Concepts and Applications 9783030374686, 3030374688

This books objective is to explore the concepts and applications related to Internet of Things with the vision to identi

3,397 519 14MB

English Pages XVII, 532 [526] Year 2020

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Internet of Things (IoT): Concepts and Applications
 9783030374686, 3030374688

Table of contents :
Preface
Acknowledgments
Contents
Contributors
Part I: Internet of Things (IoT) Architecture
Chapter 1: Foundation of IoT: An Overview
1.1 Introduction
1.2 Historical Development
1.3 Internet of Things (IoT)
1.4 Smart Object
1.4.1 Characteristics of Smart Object
1.4.2 Trends in Smart Object
1.5 Features and Challenges of IoT
1.6 IoT as an Industrial Commodity
1.7 Applications
1.8 Conclusion
References
Chapter 2: Cloud Computing for IoT
2.1 Introduction
2.2 Basic Concepts
2.2.1 Cloud Computing
2.2.2 Importance of Cloud Computing for IoT
2.3 Cloud Based IoT Architecture
2.3.1 Models
2.4 CloudIoT Applications
2.5 Cloud Platforms Available for IoT
2.5.1 Commercial IoT Platform
2.5.2 Open Source IoT Platform
2.6 Challenges
2.7 Conclusion
References
Chapter 3: Open Service Platforms for IoT
3.1 Introduction
3.2 IoT Reference Architecture
3.2.1 Sensor Layer
3.2.2 Gateway and Network Layer
3.2.3 Service Management Layer
3.2.4 Application Layer
3.3 IoT Platform Requirements
3.3.1 Service Requirements
3.3.2 Architectural Requirements
3.4 Case Study of IoT Service Platforms
3.4.1 Amazon Web Service (AWS IoT)
3.4.2 Microsoft Azure IoT
3.4.3 Google Cloud Platform
3.4.4 IBM Watson IoT
3.5 Challenges and Open Research Problems
3.6 Conclusion
References
Part II: Solutions and Enablers for IoT
Chapter 4: Resource Management Techniques for Cloud-Based IoT Environment
4.1 Introduction
4.2 Basic Concepts of IoT, Cloud and Resource Management
4.2.1 What Are the Basic Elements of the IoT Environment?
4.2.1.1 Identifiers
4.2.1.2 Sensing Devices
4.2.1.3 Communications Devices
4.2.1.4 Compute Devices
4.2.1.5 Services IoT
4.2.1.6 Semantics
4.2.2 What Are the Various IoT Architecture Frameworks?
4.2.2.1 Layer 1: Physical Devices and Controllers
4.2.2.2 Layer 2: Connectivity
4.2.2.3 Layer 3: Edge Connecting
4.2.2.4 Layer 4: Data Accumulation
4.2.2.5 Layer 5: Data Abstraction
4.2.2.6 Layer 6: The Application Layer
4.2.2.7 Layer 7: Collaboration and Processes
4.2.3 How Cloud Computing Supports IoT Infrastructure?
4.2.4 Why Resource Allocation Is Important for IoT?
4.3 Related Works
4.4 Classification of Cloud-Based IoT Resource Management Techniques
4.4.1 SLA-Aware IoT Resource Allocation
4.4.2 Context-Aware IoT Resource Allocation
4.4.3 QoS-Aware IoT Resource Allocation
4.4.4 Energy-Aware IoT Resource Allocation
4.4.5 Cost-Aware IoT Resource Allocation
4.5 Parameters of IoT Resource Management Techniques
4.5.1 What Are the Various Parameters of Resource Allocation?
4.5.2 How Much Degree of Improvement Have Been Done in These Parameters?
4.5.3 How Much Improvement Is Required in the Remaining Parameters?
4.6 Challenges and Issues
4.7 Future Directions
4.8 Conclusion
References
Chapter 5: Data Management for the Internet of Things
5.1 Introduction
5.2 Management of Data in IoT
5.2.1 Life-Cycle of Information
5.2.2 Management of Information for IoT and Data Management Systems for the Traditional Methods
5.3 Systematic Survey of the Management for the Devices of IoT and Design Primitives
5.3.1 Data Collection and Information Management Systems
5.3.1.1 Collection Strategy for Data
5.3.2 Design Elements for Database Framework
5.3.3 Preparing Elements
5.3.3.1 Access Model
5.3.3.2 Proficient Handling Procedure
5.3.3.3 Versatile Query Handling, Streamlining and Optimization
References
Chapter 6: Machine Learning for IoT Systems
6.1 Introduction
6.2 IoT Overview
6.3 Machine Learning Taxonomy
6.3.1 Supervised Learning
6.3.2 Unsupervised Learning
6.3.3 Reinforcement Learning
6.3.4 Evolutionary Computation
6.3.5 Fuzzy Logic
6.4 Machine Learning for IoT Basic Operation
6.4.1 Node Localization
6.4.2 Clustering
6.4.3 Routing
6.4.4 Data Aggregation
6.5 Machine Learning for IoT Performance Aspects
6.5.1 Congestion Control
6.5.2 Fault Detection
6.5.3 Resource Management
6.5.4 Security
6.6 Concluding Remarks
References
Chapter 7: Supervising Data Transmission Services Using Secure Cloud Based Validation and Admittance Control Mechanism
7.1 Introduction
7.2 Objectives and Challenges of Cloud Based Data Transmission Research
7.2.1 Objectives
7.2.2 Challenges
7.3 Access Control Mechanisms of Data Transmission Using Cloud Services
7.3.1 Securing Virtual Machines
7.4 Governance and Operations in Cloud Computing Environment
7.5 Security Issues for Cloud Servers and Transmission Systems
7.5.1 Security Concerns in Data/Video Transmission Through Cloud Server
7.5.2 Transportation System Requirements
7.5.3 Secure Cloud Test Setup Model
7.5.4 Performance Evaluation
7.6 Conclusions
References
Part III: IoT Challenges and Issues
Chapter 8: Tackling Jamming Attacks in IoT
8.1 Introduction
8.2 Literature Survey
8.3 Proposed Study
8.3.1 Notations Used in the Study
8.3.2 Assumptions Made in the Study
8.3.3 Creating Awareness
8.3.4 Identifying the Jammer Location
8.3.5 Choosing Appropriate Path for Delivery
8.4 Analysis of the Work
8.4.1 Energy Consumption
8.4.2 Taking Jammer into Confidence
8.4.3 Communication Overhead
8.4.4 Simulation of the Work
8.5 Conclusion
References
Chapter 9: Bioinspired Techniques for Data Security in IoT
9.1 Introduction
9.2 Data Security in IoT
9.3 Bioinspired Computing
9.3.1 Relationship Between Traditional and Bio-Inspired Data Security
9.3.2 Achieving Security in IoT Using Bioinspired Techniques
9.3.3 Types of Bioinspired Computing Algorithms
9.4 Different Approaches for Data Security in IoT Using Bioinspired Computing
9.4.1 Ant Colony Optimization (Birattari et al. 2007; Dorigo and Birattari 2011)
9.4.1.1 Routing Protocols Derived from Ant Colony Optimization for Data Security in IoT (Liu 2017)
Types of ACO Inspired Routing in WSNs
Behavior of Operation
Main Aim
Topology of Network
Probability Transition
9.4.1.2 Ant Colony Approach to Solve Travelling Salesman Problem (Beckers et al. 1992; Bolondi and Bondanza 1993)
Ant Specific Algorithm
9.4.2 Genetic Bee Colony (GBC) Algorithm (Alshamlan et al. 2015)
9.4.2.1 Bee Colony Algorithm Used for Data Security Using Routing System Protocols (Okdem et al. 2011)
Fire Evacuation Routing and Artificial Bee Colony Optimization (BCO)
Bee Colony Algorithm and Applications
Finding Fire Evacuation Route Using BCO Algorithm
Optimal Routing Solution for Fire Evacuation
9.4.2.2 Dependable Data Gathering in IoT Using Bee Colony (Najjar-Ghabel et al. 2018)
The Proposed Model
Reliable Spanning Tree-Based Data Gathering in IoT
9.4.3 Firefly Algorithm (Yang 2008)
9.4.3.1 Analysis of Performance Using Firefly Algorithm Used for the Purpose of Data Clustering (Banati and Bajaj 2013)
Analysis of Performance
Using Artificial Data Sets
Using Real World Data Sets
Percentage of Success
9.4.3.2 Security System for Image in IoT with High Performance (Alam et al. 2013; Suwetha et al. 2017)
Methodology
Image Hiding Using Firefly Algorithm
Process of Hiding Images
9.5 Conclusion
References
Chapter 10: A Chaos-Based Multi-level Dynamic Framework for Image Encryption
10.1 Introduction
10.2 Background
10.2.1 Chaos in Cryptography
10.2.2 Survey on Dynamism in Encryption
10.3 Proposed Approach
10.3.1 Description of the Proposed Framework
10.3.2 Description of Per-round Operations
10.3.3 Definition of Diffusion Stage
10.3.4 Key Description
10.4 Results
10.4.1 NPCR, UACI and Co-relation Coefficient
10.4.2 Histogram and Entropy
10.4.3 Avalanche Properties
10.4.4 NIST Statistical Test Suite for Randomness
10.4.5 Resistance Against Known/Chosen Plaintext Attacks & Differential Cryptanalysis
10.5 Conclusion & Future Scope
References
Chapter 11: Privacy Challenges and Their Solutions in IoT
11.1 Introduction
11.2 Privacy Requirements for IoT
11.3 IoT Privacy Research Analysis
11.4 Security and Privacy Concerns/Challenges
11.5 Theoretical Solutions Provided for IoT Technology
11.6 Privacy or Security Solutions from Technical and Industry Areas
11.6.1 Impact of Security in Heterogeneous Environment
11.6.2 Industrial Solutions
11.6.2.1 Security Embedded with Respect to IOT Design
11.6.2.2 Data Minimization
11.6.2.3 Transparency Among Consumers
11.7 Conclusion
References
Part IV: The IoT World of Applications
Chapter 12: Mobile Computing and IoT: Radio Spectrum Requirement for Timely and Reliable Message Delivery Over Internet of Vehicles (IoVs)
12.1 Introduction
12.2 System Model
12.2.1 MAC Layer Models
12.2.1.1 CSMA/CA Based MAC Algorithm
12.2.1.2 STDMA Based MAC Algorithm
12.2.2 PHY Layer Model
12.3 Performance Analysis
12.3.1 Safety Message Transmission (MAC-to-MAC) Delay
12.3.2 Simulation Setup
12.3.3 Simulation Settings and Assumptions
12.3.4 Results and Discussion
12.3.4.1 Probability of Message Reception Failure
12.3.4.2 Safety Message Transmission Delay
12.4 Conclusion
References
Chapter 13: Single Activity Recognition System: A Review
13.1 Introduction
13.2 Smart Home Assisted Living System
13.3 Assisted Living System for Energy Saving
13.4 Assisted Living System for Health Care
13.5 Assisted Living System for Safety and Security
13.6 Assisted Living System for Anomalous Situation Detection
13.7 Assisted Living System for Monitoring Daily Activities
13.8 Activity Recognition in Smart Home Assisted Living
13.9 Communication System in Smart Home Assisted Living
13.10 Conclusion
References
Chapter 14: Deep Learning and IoT for Agricultural Applications
14.1 Introduction
14.2 IoT in Agriculture
14.3 Deep Learning Overview
14.4 Deep Learning for Smart Agriculture: Concepts, Algorithms, Frameworks and Applications
14.4.1 Common Deep Learning Algorithms
14.4.1.1 Convolutional Neural-Network (CNN)
14.4.1.2 Recurrent Neural Networks (RNN)
14.4.1.3 Generative Adversarial Networks (GAN)
14.4.1.4 Long- Short Term Memory (LSTM)
14.4.2 Deep Learning Frameworks
14.4.2.1 TensorFlow
14.4.2.2 Caffe (Convolution Architecture for Feature Extraction)
14.4.2.3 PyTorch
14.4.2.4 Theano
14.5 Applications of Deep Learning in Agriculture
14.6 Conclusion
References
Chapter 15: IoT for Crowd Sensing and Crowd Sourcing
15.1 Introduction
15.2 Internet of Things
15.3 Crowdsourcing
15.4 Classification of Crowdsourcing
15.4.1 Collective Intelligence for Problem Solving
15.4.2 Learning paradigms
15.4.3 Open Innovation
15.4.4 New Product Development or Crowd Creation
15.4.5 Collaborative Initiative Through Crowd Funding
15.4.6 Crowd Voting
15.4.7 Crowd Curation
15.4.8 User-Generated Content (UGC)
15.4.9 Crowd Labour
15.5 Applications of Crowdsourcing in India
15.6 Crowd Sensing
15.7 Industry Initiatives
15.7.1 Challenges Faced by the Industry
15.8 Conclusion
References
Chapter 16: Smart Infrastructures
16.1 Introduction
16.2 Smart City
16.2.1 Components of smart city
16.2.1.1 Smart Parking
16.2.1.2 Smart Traffic
The Need for Smart Traffic Systems
IOT Based Smart Traffic System
Smart Lighting
Smart Hospitals
Smart Waste Disposal
IOT Based Smart Waste Disposal System
16.3 Smart Home
16.3.1 Purpose of Home Automation
16.4 Conclusion
References
Part V: IoT for Smart Cities
Chapter 17: IoT Application for Smart Cities Data Storage and Processing Based on Triangulation Method
17.1 Introduction
17.2 Overview of Related Research
17.3 Data Processing and Storing in Java NetBeans Environments
17.4 Triangulation Method for Data Storage in Smart City Space
17.5 Triangulation Method for Data Processing in Smart City Space
17.6 Discussion
17.7 Conclusion
References
Chapter 18: Intelligent Environment Protection
18.1 Introduction
18.1.1 Environment Protection
18.1.2 Preventing the Environment Degradation
18.2 Environmental Issues
18.2.1 Unauthorized Waste Dumping
18.2.2 Constitutional Nuisance
18.2.3 Unregulated Usage of Natural Resources
18.3 Monitoring of Environmental Hazards
18.4 Intelligent Environment
18.4.1 Physical Layer of the IoT Eco-System
18.4.2 Middleware Layer of the IoT Eco-System
18.4.3 Security Layer of the IoT Eco-System
18.4.4 Diagnosis Layer of the IoT Eco-System
18.5 Protection Policies
18.5.1 Regulatory Method of Environmental Policy
18.5.2 Economical Method of Environmental Policy
18.5.3 Voluntary Method of Environmental Policy
18.5.4 Education & Information Method of Environmental Policy
18.6 Conclusions
References
Chapter 19: A Decade Survey on Internet of Things in Agriculture
19.1 Introduction
19.2 Preliminary Research
19.2.1 Agriculture
19.2.1.1 Definition
19.2.1.2 Branches of Agriculture
19.2.1.3 Top 10 Agriculture Based Economies
19.2.1.4 Technologies Associated with Agriculture
19.2.2 IOT
19.2.2.1 Definition
19.2.2.2 Framework
19.2.2.3 Issues and Challenges
19.2.2.4 Applications
19.3 Literature Survey
19.3.1 Crop Cultivation
19.3.2 Livestock Production
19.3.3 Agronomics
19.3.4 Agriculture Engineering
19.4 Conclusion
References
Chapter 20: Intelligent Healthcare Solutions
20.1 Introduction
20.2 IT Platforms in Healthcare
20.2.1 Device Layer
20.2.2 Network Layer
20.2.3 Application Layer
20.3 Handling Data
20.3.1 Device Centric
20.3.2 Gateway Centric
20.3.3 Fog Centric
20.3.4 Cloud Centric
20.3.5 Intelligent Approaches Towards Data Processing
20.3.6 Data Validation
20.4 Personalised Healthcare Systems
20.5 Conclusion
References
Chapter 21: Smart Car – Accident Detection and Notification Using Amazon Alexa
21.1 Introduction
21.2 Related Work
21.3 Proposed System
21.4 Proposed Algorithm
21.5 Experimental Issues
21.6 Results and Analysis
21.6.1 Simulation for Moving Car Accident
21.7 Conclusion
References
Chapter 22: Prioritisation of Challenges Towards Development of Smart Manufacturing Using BWM Method
22.1 Introduction
22.2 Literature Review
22.2.1 Overview of Smart Manufacturing
22.2.2 Smart Manufacturing Related Studies
22.3 Research Methodology
22.4 Result
22.4.1 Identification of Challenges in Smart Manufacturing
22.4.2 Prioritisation of the Challenges Towards Development of Smart Manufacturing
22.5 Discussion on Results
22.6 Conclusion, Limitations and Future Scope
References
Part VI: Next Generation Smart Applications
Chapter 23: Surveillance of Type –I & II Diabetic Subjects on Physical Characteristics: IoT and Big Data Perspective in Healthcare @NCR, India
23.1 Introduction
23.1.1 Role of Big Data in Healthcare
23.1.1.1 Big Data Characteristics and Its Benefits in Healthcare
23.1.1.2 The Future of Healthcare Big Data
23.1.1.3 Role of Big Data in IoT
23.1.1.4 Big Data Tools
23.1.1.5 Big Data Security
23.1.2 IoT
23.1.2.1 Origin of IoT
23.1.2.2 Applications of IoT
23.1.2.3 Applying Internet of Things (IoT) for Healthcare
23.1.2.4 Healthcare Applications of IoT
23.1.2.5 Critical Issues and Challenges of IoT in Healthcare
23.1.2.6 Examples of IoT Services in Healthcare
23.1.3 Artificial Intelligence (AI) & Machine Learning (ML)
23.1.4 Tension Type and Chronic Headache
23.1.4.1 TTH Treatment
23.1.4.2 TTH Preventions
23.1.5 Diabetes Mellitus and Its Types
23.1.5.1 Facts for Diabetes Cause Tension Type Headache
23.1.5.2 Diabetes and Headaches
23.1.6 Hypoglycemia (Low Blood Sugar)
23.1.6.1 Hypoglycemia and Diabetes
23.1.6.2 Hypoglycemia and Headaches
23.1.6.3 Hyperglycemia and Its Causes
23.1.6.4 Hyperglycemia and Headaches
23.1.6.5 The Cases When to See a Doctor
23.1.7 Depression
23.1.7.1 Depression: Sign and Symptoms
23.1.7.2 Depression Causes
23.1.7.3 Depression Treatment
23.1.8 Obesity
23.1.9 CAD
23.1.9.1 CAD Causes
23.1.9.2 CAD Symptoms
23.1.9.3 CAD Treatment
23.1.10 Insulin
23.2 Literature Survey
23.3 Our Experimental Results, Interpretation and Discussion
23.3.1 Experimental Setup
23.3.2 About the Study and Analysis
23.4 Novelty in Presented Work
23.5 Future Scopes, Limitations and Possible Applications
23.6 Recommendations and Future Considerations
23.7 Conclusions
References
Chapter 24: Monitoring System Based in Wireless Sensor Network for Precision Agriculture
24.1 Introduction
24.2 Related Work
24.2.1 Agricultural Monitoring System
24.3 Wireless Sensor Network for Precision Agriculture Application
24.3.1 System Overview
24.3.2 Routing Protocol
24.4 Experimental Results and Discussion
24.4.1 Packet Delivery Ratio: PDR
24.4.2 Average Energy Consumption per Node
24.4.3 Network Lifetime
24.5 Conclusion
References
Chapter 25: Securing E-Health IoT Data on Cloud Systems Using Novel Extended Role Based Access Control Model
25.1 Introduction to Internet of Things
25.1.1 Internet of Things
25.1.2 Key Fundamentals of Internet of Things
25.1.3 Architecture of IoT
25.1.4 Standards of IoT Applications
25.2 Role of IoT in Health Care Systems
25.3 Background
25.4 Access Control and Authorization Using Role Based Access Control (RBAC)
25.4.1 Rules for Defining RBAC
25.4.2 RBAC Reference Models
25.5 Security Healthcare Model Based on Extended RBAC
25.5.1 Extended Role Based Model (ERBAC)
25.5.2 Proposed Security Model for Storage of Medical IoT Data Using ERBAC
25.6 Conclusion
References
Chapter 26: An Efficient Approach towards Enhancing the Performance of m-Health Using Sensor Networks and Cloud Technologies
26.1 Introduction
26.2 Literature Review
26.3 The Pillars and Paradigms of Mobile Health Systems
26.4 Proposed Architecture of Mobile Health System
26.4.1 Awareness Layer
26.4.2 Middleware and Application Programme Interface Layer
26.4.3 E-Mobile Health Application and Service Layer
26.5 Results and Discussions
26.5.1 Services Provided by Proposed Cloud and IoT Based Healthcare System
26.5.1.1 Cloud Data Management and Storage Services
26.5.1.2 Hospital Services
26.5.1.3 Emergency and Urgent Response Services
26.5.1.4 Online Health Advice and Necessary Action Services
26.5.1.5 Online Patient Monitoring Services
26.5.2 Online Monitoring the Performance of Cloud and IoT Based Healthcare Systems
26.6 Conclusions
References
Chapter 27: Future Internet of Things (IOT) from Cloud Perspective: Aspects, Applications and Challenges
27.1 Introduction
27.2 Background
27.2.1 Understanding IoT
27.2.2 IoT and Its Relation with Cloud
27.3 Application of IoT
27.3.1 Smart Home
27.3.2 Wearables Technologies
27.3.3 Smart City
27.3.4 Smart Grids
27.3.5 Industrial Internet
27.3.6 Connected Car
27.3.7 Connected Health (Digital Health/Tele Health/Tele Medicine)
27.3.8 Smart Retail
27.3.9 Smart Supply Chain
27.3.10 Smart Farming
27.3.11 Smart Factories
27.3.12 Smart Food Industry
27.4 Future of IoT
27.4.1 IoT Network in Future
27.5 Conclusion
References

Citation preview

Mansaf Alam Kashish Ara Shakil Samiya Khan  Editors

Internet of Things (IoT) Concepts and Applications

Internet of Things (IoT)

Mansaf Alam  •  Kashish Ara Shakil Samiya Khan Editors

Internet of Things (IoT) Concepts and Applications

Editors Mansaf Alam Department of Computer Science Jamia Millia Islamia New Delhi, India Samiya Khan Department of Computer Science Jamia Millia Islamia New Delhi, India

Kashish Ara Shakil College of Computer & Information Science Princess Nourah bint Abdulrahman University Riyadh, Saudi Arabia

ISBN 978-3-030-37467-9    ISBN 978-3-030-37468-6 (eBook) https://doi.org/10.1007/978-3-030-37468-6 © Springer Nature Switzerland AG 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

The emergence and rise of the Internet of Things (IoT) promises to revolutionize the way humans lead their lives by connecting all the devices that we possibly use to a common network, the Internet. Some of these devices include sensors, home appliances, health monitoring devices, and any device that produces quantifiable data. This concept has given birth to unlimited application areas that use the interactions between humans and devices to build inferences and make predictions for improving the quality of life and optimizing the use of existing resources. A recent survey of McKinsey has shown that the number of deployed IoT devices is expected to reach one trillion by 2025. Moreover, the economic impact of this technology on the world economy can be assessed from the fact that it is expected to capture 11% of the same by the end of 2025. With that said, Internet of Things, as a technological paradigm, suffers from several challenges that must be addressed before its vision can be realized. Some of the evident challenges include identification and development of architectures that can meet the scalability requirements of the IoT ecosystem. Data and resource management along with the need to maintain the security and privacy of the system also remain prevalent, in addition to several others. All the devices that fall under the umbrella of the Internet of Things are commonly referred to as smart devices. Therefore, the concept of IoT acquires and integrates data from these smart devices. Furthermore, this data is stored and processed to generate useful analytics. The vision of IoT is to visualize the world as a collection of connected entities. The most obvious application of such a paradigm is smart city as the connected nature of the city is expected to make resource management simpler. Among other applications of IoT, healthcare and infrastructure management are frontrunners because of their need for real-time solutions in critical scenarios. In addition, there are a plethora of applications that can be developed and commercialized for human use. The objective of this book is to explore the concepts and applications related to the Internet of Things with the vision to identify and address existing challenges. Besides this, it shall also provide future research directions in this domain. This book is meant for students, practitioners, industry professionals, researchers, and v

vi

Preface

faculty working in the field of Internet of Things and its integration with other technologies to develop integrated, comprehensive solutions to real-life problems. Part I introduces the basic concept of the Internet of Things and provide an insight into other parts of the book and what they are expected to cover. In order to implement Internet of Things solutions, the architecture must support the specific requirements of such applications. These requirements include scalability, transition from closed platforms to open platforms, and designing of protocols for interaction at different levels. This part also covers the architectural issues and available solutions related to IoT. Once the architectural issues are discussed and elaborated upon, the next part is expected to cover the solutions available in this domain for elemental IoT processes like data and device management. The heart of the IoT ecosystem is a smart device and the programming framework that uses the data made available by the smart device to create useful insights. Part II covers components that allow development of solutions and applications using the IoT paradigm. The concept of the Internet of Things has just hit the shore. There are a number of challenges and limitations that need to be mitigated to make this technology workable and usable across diverse domains. Some of the identified challenges include security, robustness, reliability, privacy, identity management, and designing of management policies to ensure smooth functioning. Part III covers the different aspects of IoT challenges and devised solutions for the same. The emergence and growing popularity of the Internet of Things has given rise to the identification of many areas where it can be put to use. Some of the obvious applications of this paradigm include social computing, mobile computing, crowd sensing, and crowd sourcing. Part IV includes chapters that have implemented IoT to create applications for these domains. Smart cities is the most popular application of IoT and uses the same in conjunction with other technologies like cloud computing and big data. Smart cities is a large domain of applications that include domains like healthcare, logistics, manufacturing, and agriculture, in addition to many others. Part V includes chapters that explore the different facets of smart cities and solutions created for the same. Next-generation smart applications make use of many state-of-the-art technologies like machine learning, computer vision, and artificial intelligence in conjunction with the Internet of Things (IoT). Moreover, deep learning has also found many integrative applications with IoT. These applications are popularly termed as cognitive IoT applications and promise to serve a wide range of areas and domains like healthcare, logistics, smart cities, and supply chain, in addition to many others. Part VI elaborates on the concepts and applications related to cognitive IoT analytics and its applications. New Delhi, India Riyadh, Saudi Arabia  New Delhi, India 

Mansaf Alam Kashish Ara Shakil Samiya Khan

Acknowledgments

The making of this book is a long journey that required a lot of hard work, patience, and persistence. We wish to express our heartfelt gratitude to our families, friends, colleagues, and well-wishers for their endless support throughout this journey. We would particularly like to express our gratitude to Mr. A.  P. Siddiqui, Registrar, Jamia Millia Islamia, New Delhi, for his constant encouragement. We would also like to thank Prof. Haroon Sajjad, Dr. Arshad Khan, Dr. Israr Ahmad, Dr. Khalid Raza, and Mr. Abdul Aziz for their unconditional support. Besides, we owe a deep sense of appreciation to other members of our research lab for their cooperation. Lastly, we wish to acknowledge and appreciate the Springer team for their continuous support throughout the entire process of publication. Our gratitude is extended to the readers, who gave us their trust, and we hope this work guides and inspires them. Mansaf Alam Kashish Ara Shakil Samiya Khan

vii

Contents

Part I Internet of Things (IoT) Architecture 1 Foundation of IoT: An Overview������������������������������������������������������������    3 Zaheeruddin and Hina Gupta 2 Cloud Computing for IoT�����������������������������������������������������������������������   25 Himani Tyagi and Rajendra Kumar 3 Open Service Platforms for IoT��������������������������������������������������������������   43 Preeti Agarwal and Mansaf Alam Part II Solutions and Enablers for IoT 4 Resource Management Techniques for Cloud-Based IoT Environment��������������������������������������������������������������������������������������   63 Syed Arshad Ali, Manzoor Ansari, and Mansaf Alam 5 Data Management for the Internet of Things����������������������������������������   89 Amrit Sahani, Ranjit Kumar, Suchismita Chinara, Anjali Kumari, and Bina Patro 6 Machine Learning for IoT Systems��������������������������������������������������������  105 Ahmed Khattab and Nouran Youssry 7 Supervising Data Transmission Services Using Secure Cloud Based Validation and Admittance Control Mechanism������������������������  129 Kamta Nath Mishra Part III IoT Challenges and Issues 8 Tackling Jamming Attacks in IoT������������������������������������������������������    153 N. Ambika 9 Bioinspired Techniques for Data Security in IoT����������������������������������  167 S. R. Mani Sekhar, G. M. Siddesh, Anjaneya Tiwari, and Ankit Anand ix

x

Contents

10 A Chaos-Based Multi-level Dynamic Framework for Image Encryption������������������������������������������������������������������������������  189 Sakshi Dhall, Saibal K. Pal, and Kapil Sharma 11 Privacy Challenges and Their Solutions in IoT������������������������������������  219 Nabeela Hasan, Akshay Chamoli, and Mansaf Alam Part IV The IoT World of Applications 12 Mobile Computing and IoT: Radio Spectrum Requirement for Timely and Reliable Message Delivery Over Internet of Vehicles (IoVs)��������������������������������������������������������������������������������������  235 Elias Eze, Paul Sant, Sijing Zhang, Xiaohua Feng, Mitul Shukla, Joy Eze, and Enjie Liu 13 Single Activity Recognition System: A Review��������������������������������������  257 P. K. Nizar Banu and R. Kavitha 14 Deep Learning and IoT for Agricultural Applications ������������������������  273 Disha Garg and Mansaf Alam 15 IoT for Crowd Sensing and Crowd Sourcing����������������������������������������  285 Vinita Sharma 16 Smart Infrastructures������������������������������������������������������������������������������  301 Zameer Fatima, Lakshita Bhargava, and Alok Kumar Part V IoT for Smart Cities 17 IoT Application for Smart Cities Data Storage and Processing Based on Triangulation Method ������������������������������������������������������������  317 Muzafer Saračević, Šemsudin Plojović, and Senad Bušatlić 18 Intelligent Environment Protection��������������������������������������������������������  335 Subha P. Eswaran 19 A Decade Survey on Internet of Things in Agriculture������������������������  351 Ummesalma M, Rachana Subbaiah M, and Srinivas Narasegouda 20 Intelligent Healthcare Solutions ������������������������������������������������������������  371 Salman Basheer Ahmed and B. M. Jabarullah 21 Smart Car – Accident Detection and Notification Using Amazon Alexa��������������������������������������������������������������������������������  391 Lakshay Grover, V. B. Kirubanand, and Joy Paulose 22 Prioritisation of Challenges Towards Development of Smart Manufacturing Using BWM Method ������������������������������������  409 Shahbaz Khan, Mohd Imran Khan, and Abid Haleem

Contents

xi

Part VI Next Generation Smart Applications 23 Surveillance of Type –I & II Diabetic Subjects on Physical Characteristics: IoT and Big Data Perspective in Healthcare @NCR, India��������������������������������������������������������������������  429 Rohit Rastogi, D. K. Chaturvedi, Santosh Satya, Navneet Arora, Parul Singhal, and Mayank Gupta 24 Monitoring System Based in Wireless Sensor Network for Precision Agriculture ������������������������������������������������������������������������  461 Fekher Khelifi 25 Securing E-Health IoT Data on Cloud Systems Using Novel Extended Role Based Access Control Model������������������������������  473 Mamoon Rashid, Shabir Ahmad Parah, Aabid Rashid Wani, and Sachin Kumar Gupta 26 An Efficient Approach towards Enhancing the Performance of m-Health Using Sensor Networks and Cloud Technologies������������  491 Kamta Nath Mishra 27 Future Internet of Things (IOT) from Cloud Perspective: Aspects, Applications and Challenges����������������������������������������������������  515 Nahid Sami, Tabish Mufti, Shahab Saquib Sohail, Jamshed Siddiqui, Deepak Kumar, and Neha

Contributors

Preeti  Agarwal  Department of Computer Science, Jamia Millia Islamia, New Delhi, India Salman  Basheer  Ahmed  Department of Computer Engineering, Faculty of Engineering and Technology, New Delhi, India Mansaf  Alam  Department of Computer Science, Jamia Millia Islamia, New Delhi, India Syed  Arshad  Ali  Department of Computer Science, Jamia Millia Islamia, New Delhi, India N.  Ambika  Department Bangalore, India

of

Computer

Applications,

SSMRV

College,

Ankit  Anand  Department of Information Science & Engineering, Ramaiah Institute of Technology, Bangalore, Karnataka, India Manzoor  Ansari  Department of Computer Science, Jamia Millia Islamia, New Delhi, India Navneet Arora  Department of ME, IIT- Roorkee, Roorkee, Uttarakhand, India Lakshita  Bhargava  Computer Science & Engineering Department, Maharaja Agrasen Institute of Technology, Guru Gobind Singh Indraprastha University, New Delhi, Delhi, India Senad  Bušatlić  International University of Sarajevo, Sarajevo, Bosnia and Herzegovina Akshay  Chamoli  Department of Computer Science, Jamia Millia Islamia, New Delhi, India D. K. Chaturvedi  Department of Electrical Engineering, DEI-Agra, Agra, Uttar Pradesh, India

xiii

xiv

Contributors

Suchismita Chinara  National Institute of Technology, Rourkela, Odisha, India Sakshi Dhall  Department of Mathematics, Jamia Millia Islamia, New Delhi, India Department of Computer Science & Engineering, Delhi Technological University, New Delhi, India Subha P. Eswaran  CRL, Bharat Electronics Limited, Bangalore, Karnataka, India Elias  Eze  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK Joy  Eze  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK Zameer Fatima  Computer Science & Engineering Department, Maharaja Agrasen Institute of Technology, Guru Gobind Singh Indraprastha University, New Delhi, Delhi, India Xiaohua  Feng  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK Disha  Garg  Department of Computer Science, Jamia Millia Islamia, New Delhi, India Lakshay  Grover  Department Bangalore, India

of

Computer

Science,

Christ

University,

Hina  Gupta  Department of Electrical Engineering, Jamia Millia Islamia, New Delhi, India Mayank Gupta  IT Consultant, TCS, Noida, Uttar Pradesh, India Sachin Kumar Gupta  Department of Electronics & Communication, Shri Mata Vaishno Devi University, Katra, Jammu, India Abid Haleem  Department of Mechanical Engineering, Faculty of Engineering and Technology, Jamia Millia Islamia, New Delhi, India Nabeela  Hasan  Department of Computer Science, Jamia Millia Islamia, New Delhi, India B. M. Jabarullah  Pusa Institute of Technology, New Delhi, India R. Kavitha  Department of Computer Science, CHRIST (Deemed to be University), Bangalore, India Mohd  Imran  Khan  Department of Mechanical Engineering, Faculty of Engineering and Technology, Jamia Millia Islamia, New Delhi, India Shahbaz  Khan  Department of Mechanical Engineering, Faculty of Engineering and Technology, Jamia Millia Islamia, New Delhi, India

Contributors

xv

Ahmed  Khattab  Electronics and Electrical Communications Engineering Department, Cairo University, Giza, Egypt Fekher  Khelifi  Laboratory of Electronics and Microelectronics, University of Monastir, Monastir, Tunisia V.  B.  Kirubanand  Department of Computer Science, Christ University, Bangalore, India Alok  Kumar  Computer Science & Engineering Department, Maharaja Agrasen Institute of Technology, Guru Gobind Singh Indraprastha University, New Delhi, Delhi, India Deepak  Kumar  Amity Institute of Information Technology, Amity University, Noida, India Rajendra Kumar  Jamia Millia Islamia, New Delhi, India Ranjit Kumar  National Institute of Technology, Rourkela, Odisha, India Anjali Kumari  College of Engineering Roorkee, Roorkee, Uttarakhand, India Enjie  Liu  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK S. R. Mani Sekhar  Department of Information Science & Engineering, Ramaiah Institute of Technology, Bangalore, Karnataka, India Kamta  Nath  Mishra  Department of Computer Science & Engineering, Birla Institute of Technology, Ranchi, Jharkhand, India Tabish Mufti  Department of Computer Applications, Faculty of Computer Science and System Studies, Mewar University, Chittorgarh, Rajasthan, India Srinivas  Narasegouda  Department of Computer Science, Jyothi Nivas College (Autonomous), Bengaluru, Karnataka, India Neha  Department of Computer Science and Engineering, School of Engineering Sciences and Technology, Jamia Hamdard, New Delhi, India P.  K.  Nizar  Banu  Department of Computer Science, CHRIST (Deemed to be University), Bangalore, India Saibal K. Pal  Scientific Analysis Group, DRDO, New Delhi, India Shabir Ahmad Parah  Department of Electronics and Instrumentation Technology, University of Kashmir, Srinagar, India Bina Patro  University of Berhampur, Berhampur, Odisha, India Joy Paulose  Department of Computer Science, Christ University, Bangalore, India

xvi

Contributors

Šemsudin Plojović  Department of Computer Sciences, University of Novi Pazar, Novi Pazar, Serbia Rachana Subbaiah M  Department of Computer Science, CHRIST (Deemed to be University), Bengaluru, Karnataka, India Mamoon Rashid  School of Computer Science & Engineering, Lovely Professional University, Jalandhar, India Rohit Rastogi  DEI Agra, Agra, India ABESEC, Ghaziabad, Uttar Pradesh, India Amrit Sahani  Siskha ‘O’ Anusandhan University, Bhubaneswar, Odisha, India Institute of Technical Education and Research, Bhubaneswar, Odisha, India Nahid  Sami  Department of Computer Science and Engineering, School of Engineering Sciences and Technology, Jamia Hamdard, New Delhi, India Paul  Sant  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK Muzafer Saračević  Department of Computer Sciences, University of Novi Pazar, Novi Pazar, Serbia Santosh Satya  Department of Rural Development, IIT-Delhi, New Delhi, India Kapil  Sharma  Department of Information Technology, Delhi Technological University, New Delhi, India Vinita Sharma  New Delhi Institute of Management, New Delhi, India Mitul  Shukla  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK G.  M.  Siddesh  Department of Information Science & Engineering, Ramaiah Institute of Technology, Bangalore, Karnataka, India Jamshed Siddiqui  Department of Computer Science, Aligarh Muslim University, Aligarh, India Parul Singhal  DEI Agra, Agra, India ABESEC, Ghaziabad, Uttar Pradesh, India Shahab Saquib Sohail  Department of Computer Science and Engineering, School of Engineering Sciences and Technology, Jamia Hamdard, New Delhi, India Anjaneya  Tiwari  Department of Information Science & Engineering, Ramaiah Institute of Technology, Bangalore, Karnataka, India Himani Tyagi  Jamia Millia Islamia, New Delhi, India Ummesalma  M  Department of Computer Science, CHRIST (Deemed to be University), Bengaluru, Karnataka, India

Contributors

xvii

Aabid  Rashid  Wani  Department of Electronics & Communication, Shri Mata Vaishno Devi University, Katra, Jammu, India Nouran  Youssry  Electronics and Electrical Communications Engineering Department, Cairo University, Giza, Egypt Zaheeruddin  Department of Electrical Engineering, Jamia Millia Islamia, New Delhi, India Sijing  Zhang  Department of Computer Science and Technology, University of Bedfordshire, Luton, UK

Part I

Internet of Things (IoT) Architecture

Chapter 1

Foundation of IoT: An Overview Zaheeruddin and Hina Gupta

Abstract  The advancement in technology has depreciated the size of the gadget by leaps and bounds during the last few decades. The usage and dependence of human beings on electronic gadgets has grown exponentially with the passage of time. The substantial scaling of such kind of devices gave birth to a new technology called ‘Internet of Things’ (IoT) in the year 1999. The concept of IoT adheres to the need of round the clock and ubiquitous connectivity. It provides means for establishing interconnectivity and interactions between such devices. IoT has carved out the path for new innovations leading to novel type of communication among humans and things. The interactions would enable the realization of services utilizing these resources for improving the quality of life. This chapter is an attempt in this direction with the basic introduction of IoT, basic architectures, challenges and the plethora of services provided by IoT. Keywords  Internet of Things (IoT) · Smart objects · IoT architecture

1.1  Introduction The modern living trends of the society have turned out to be addictive to 24 × 7 connectivity. In order to provide this round the clock connectivity, the notion of ‘Internet of Things (IoT)’ came into existence after a gap of almost four decades following the invention of Internet. Internet can be defined as a massive pool of applications and protocols built on top of network of computers (Cohen-Almagor 2011). In today’s era, ubiquitous computing and seamless connectivity is no more a challenge. Earlier the concept of connectivity was restricted to the interactions between humans and machines. But gradually with the advancement in technology of connectivity, today one can relate connectivity between anything as in IoT. The most important feature of IoT is that it can transform an object into smart object by providing sensing, actuating, computing and communicating capability to the object Zaheeruddin · H. Gupta (*) Department of Electrical Engineering, Jamia Millia Islamia, New Delhi, India e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_1

3

4

Zaheeruddin and H. Gupta

(Sadiku et al. 2018). The huge workspace for IoT is fueled by the smooth integration of the technologies like digital electronics, microelectromechanical system and nanotechnology etc. (Bello and Zeadally 2015). The functionality of the objects and technologies together provide an environment where the sharing of the information can be carried out across the varied platforms. IoT is carving out its path in all the domains and is on the verge of becoming Internet of Everything (Miorandi et al. 2012). The term “Internet of Things” was first coined by Kevin Ashton with contextual reference to supply chain management in the year 1999 (Kevin 2010). The term was further redefined by various researchers to include applications like transport, mining, healthcare, and utilities etc. The tenure which marks the emergence of IoT is stated to be during 2008–2009. In this period, the human populace began to be eclipsed by network allied devices. With the passage of time the number of linked “things”, which included human and gadgets showed exponential escalation. This carved out the necessity for “Internet of Things”. IoT can be defined as the usage of internet that bridges the gap between varied services and things including humans (Singh et al. 2014). As per the conceptual framework of 2020, IoT can be expressed in terms of mathematical language as (Alhafidh and Allen 2016).

IoT  Sensor  Network  Data  Services

IoT can be treated as the extension of the contemporary internet services to enclose each existing object. These objects can be living or non living thing which can appeal for a service or offer a service. This global connectivity has been possible by embedding sensors, actuators, microcontrollers etc. in the things, which facilitates the Internet to become more persistent. The miniaturization and advancements in the field of electronics and computers has made the materialization of IoT visible in many fields during the last few years. Some major applications of IoT are depicted in Fig. 1.1.

1.2  Historical Development Communication has been an integral part of our day to day life. It can be defined as the transmission of information from one entity or object (sender) to another entity or object (receiver) with the help of some medium. In the early age, smoke signals, drum beats, chirping of birds, waving of handkerchiefs or flags were used as means of communication. Some other modes of communication were pigeons, hydraulic semaphore system, chain of beacons at the hilltops etc. Later electric telegraph came into existence in 1816 followed by the communication of audio messages using telephone in 1870s. In order to improve the range of communication, the invention of radio waves was made in 1880s. With the novel inventions of different modes of communication, the use of radio and televisions for communication was termed as ‘Telecommunication’ (Cohen-Almagor 2011; Leiner et  al. 2009). The word

1  Foundation of IoT: An Overview

5

Fig. 1.1 Domains supported by IoT. (Alhafidh and Allen 2016) Healthcare Smart Home Smart City

Waste Management

Aviation and Aerospace

Smart Grid

Transportation

Mining

Media

Telecommunication consists of two words, tele means ‘from far distances’ and communication means ‘to share information’. In other words it means communication at a distance. The next innovation in this field highlighted in the late 1890s in the form of Fax (facsimile). This mode of communication helped in the transmission of scanned printed material over the telephone connected to printers or some other input device. This was also called telecopying. It worked by scanning the original document with the fax machine. The machine processed the contents of both image and text in the form of a single fixed graphic image. This image was converted into a bitmap and transmitted in the form of audio-frequency tones through the telephone system. The receiving fax machine interpreted the tone and reconstructed the image by printing it on the paper. In 1940, the concept of a better mode of communication came into existence, where some calculations were done on one computer and the result was displayed on the other remote computer. This concept of a main computer and the other connected remote computers became popular in 1950s. Later in 1960s, researchers were on toes and investigated a new technology called packet switching. Packet switching breaks the data into small chunks and sends it to different computers. This differed from the previous means of communication as it would not pass the data through a centralized mainframe system. However, it was not until the personal motive of Department of Defense’s Advanced Research Projects Agency (ARPAnet) that the concept of ‘networking’ came into existence (Leiner et  al. 2009). In this project, a network of military computers was formed for communication in order to survive a disastrous nuclear attack. This small network of defense computers laid the first stone of the world of INTERNET. In the next phase of development around 1970s, protocols were designed to connect devices to the network of computers. This protocol was termed as the Transmission Control Protocol and Internet Protocol (TCP/IP). This protocol

6

Zaheeruddin and H. Gupta

Fig. 1.2  Evolutionary phases of the internet. (Perera et al. 2014)

offered standards stating how the transmission could be carried out in a network of computers. The paradigm then shifted towards the concept of network of networks. Finally, the term ‘Internet’ was coined by Vinton Cerf, Yogen Dalal and Carl Sunshine in 1974 (Cohen-Almagor 2011). Internet can be interpreted as an union of two terms ‘Inter’ and ‘net’, meaning network between devices. Internet can be defined as the global connection of computer networks that transmits data using TCP/IP. The popularity of internet grew slowly and steadily as it digitized access, digitized business, digitized world and finally is digitizing everything including people, process and things as shown in Fig. 1.2.

1.3  Internet of Things (IoT) Internet of Things endeavors to provide an epoch wherein digital and physical entities will be able to seamlessly communicate to provide a new domain of applications and services. Many applications and features we use these days prove as an ascending wave in the computing era. In the imminent time, the Internet will have its existence as flawless connection of objects and network. IoT can well be explained by a Venn diagram as shown in Fig. 1.3, where two different sets namely Information Technology (IT) and Operational Technology (OT) unite to form Internet of Things. The IT domain consists of ‘things ‘like servers, databases and applications. These ‘things’ run on the network and are controlled by IT. IT assists secure connectivity of the data and gadgets for smooth flow within the vicinity of an organization. On the other hand, OT is generally concerned with the industrial work and contains ‘things’ like sensors and devices connected to the machines or some other equipment. It supervises devices and processes on physical systems (Perera et al. 2014). These systems incorporate industries, roadways systems, production s­ ervices etc. Prior to the existence of IoT, the concept of IT and OT were considered to be

7

1  Foundation of IoT: An Overview Fig. 1.3  Venn diagram

INFORMATION TECHNOLOGY

OPERATIONAL TECHNOLOGY

INTERNET OF THINGS (IoT)

poles apart that worked independently and had little requirement to interact with each other. IoT has changed this paradigm to some extent and is still working in order to collaborate IT and OT into a single concept. The concept of IoT focuses on an interconnected world in which every “thing” is connected to any “thing”. The obligatory part of IoT is to provide smart association with the in-use network and context-aware computations using network assets (Singh et  al. 2014). Some requisites for achieving the stated motives are: (a) a mutual discernment of the circumstances of the gadgets and their associated users, (b) the investigative tools that seek for autonomous and smart actions, and (c) software architectures and pervasive communication networks to process and convey the contextual information to where it is relevant. Internet is the actual backbone that has carved out the path for ubiquitous computing. It enables inter device communication across the globe. The two basic pillars for the growth of ubiquitous computing are: Internet of Things and Cloud Computing. Cloud computing provides reliable service by providing virtual storage and computing zone for processing. On the other hand, Internet of Things provides a seamless connectivity between objects using RFID, which provides a unique code to each and every object (Chemudupati et al. 2012). All the work carried out in the background of IoT and cloud computing are completely hidden from the user. The smart environment consisting of the sensor-actuator-internet framework has given rise to a common operating picture (COP), where the generated data and information is efficiently shared across varied platforms and applications. Internet of Things is a shift towards ubiquitous computing evolved due to the presence of Wi-Fi and ad hoc wireless networking. However, for the successful and widespread use of Internet of Things, networking exemplar has to step out of its boundary of mobile computing. There are ample of technologies that exist and some that are about to spring up to interconnect the devices. The foremost goal of Internet of things is to make the environment intelligent. The existence and functioning of IoT environment is possible due to the concept of digital electronics, microelectromechanical system and nanotechnology etc. All the said technologies converge to provide the ability of

8

Zaheeruddin and H. Gupta

capturing data, providing connectivity, computing and storing data for usage. The sensors, actuators, RFID, Bluetooth and Wireless Sensor Network etc. are embedded into physical items to turn them into ‘SMART’ objects thus leveraging the connectivity amongst the actual and implicit realms (Yassein et  al. 2017). Context related computations and well tuned connectivity with existing networks are the prime requisites for the IoT. As humans use Internet to communicate, the “smart things” use the Internet of things to communicate. Eventually, the quantity and quality of data collected and analyzed to land up to a solution determine the smartness of the object.

1.4  Smart Object In the late 1980s, tremendous efforts were made by the researchers to use technology as a bridge for human-to-human communication. As a result, the concept of ubiquitous computing came into existence. The major motive of ubiquitous computing was to embed technology in order to cater the everyday needs of public. In today’s era of smart phones and other handheld or pocket devices, the environment has turned out to be more interactive and informative. The focus of the interaction has shifted to smart communication. A communication is smart if it is between anything that is capable of judging its context and state. Smart communication is only possible if the environment is also smart. According to Mark Weiser, smart environment is an environment containing heterogeneous combination of objects embedded with sensors, actuators, computational capability, display area which are seamlessly connected through a network (Mühlhäuser and Gurevych 2010; Bohn et al. 2003). The implications that came out from the early studies conducted by Weiser were that any physical object is smart if it acts as the source of digital information. Marcelo Kallman and Daniel Thalmann introduced the concept of smart object (Kallmann and Thalmann 1999). They termed an object as smart object if it is capable of describing its own possible interactions. These objects have the ability of sensing, computing and communicating wirelessly both in short distance and long distance. The resemblance of smart objects and IoT can be understood by the fact that the basic building block of human body is cell. Each cell performs its operation and assists in the good functioning of the body. In a similar manner, smart objects act as the basic building block of IoT which combine to form a smart network. Each smart object assists in the formation of a network where a lot of data or information can be gathered, processed and analyzed. With strong affirmation it can be said that the real strength of IoT comes from the transformation of an isolated ordinary object into interactive smart object. The synonyms widely used for smart object are IoT device, smart device, intelligent device, intelligent node etc.

1  Foundation of IoT: An Overview

9

1.4.1  Characteristics of Smart Object Any object or device can be categorized as smart object if it possesses certain characteristics. The Smart Object consists of Sensors/Actuators, Memory, Communication Device, Power Source, and Processing Unit as shown in Fig. 1.4. The detailed description of each component is given below: 1. Sensors: As the name implies, sensors are used for sensing. A human body has five sense organs. The sense organs are responsible for interacting with the environment. The sensing performed by sense organs may be visual sensing, physical sensing or audio sensing. As soon as something is sensed, a message is send to the brain which in turn takes some decisions. In a similar manner, sensors embedded in the device helps the smart object in sensing and measuring the changes in the environment. This physical quantity when measured is converted into digital representation. This digital representation is then passed onto some other computational unit where the transformation of data is done so that it can be used by other devices or humans as shown in Fig. 1.5 (El Jai and Pritchard 2007; Lannacci 2018; Pinelis 2017). A lot of sensors are available to measure various quantity and quality of the physical and virtual things in the environment. 2. Actuators: The sensors are used for sensing data. The data is collected from various sources and devices. This data is then processed, analyzed and needs to be used to produce some productive result. This sensed data now triggers the actuators. The actuators are complements to sensors as shown in Fig.  1.5. It receives control signal and produces some response to the physical

Fig. 1.4  Components of a Smart Object

10

Zaheeruddin and H. Gupta

Fig. 1.5  Relationship between sensor and actuator

world (Anjanappa et al. 2002). In short we can say that sensors sense and send whereas actuators act and activate. A smart object may either have sensor or actuator or both and their number can vary from zero to multiple values depending on the requirement. The relationship between sensors and actuators can be understood with a simple example. Suppose there is a water tank marked with a level. The water pump should be switched on as soon as the water reaches below the marked level. In order to design such a system both sensor and actuator will be employed. Sensor will sense the level of water in the tank. When the water level reaches the marked level, a notification message will be sent to the controller embedded into the system. At this point of time the controller will trigger the actuator to turn on the water pump. In short we can say that actuators are triggering device that converts energy into motions (Alhafidh and Allen 2016; El Jai and Pritchard 2007; Lannacci 2018). Actuators can generate rotary, linear or oscillatory motion. The actuators can be classified on the basis of various parameters. In the context of IoT, switching on and off another device or equipment by the application of force is handled by actuator. IoT is an amalgamation of not only sensing and processing the data but also triggering varied devices into operation on the basis of the dynamics of data. Sensors and actuators that complement each other in terms of functionality work in collaboration to attain the maximum benefit from Internet of Things. 3 . Memory: The interaction of smart object is not only limited to physical world objects but it also includes interaction with the virtual objects. A smart object

1  Foundation of IoT: An Overview

11

may have a physical existence and physical properties such as size, shape etc. or a virtual existence in the form of software objects. In both the cases, the smart objects should have a unique identifier. This identifier would be an address which would play a huge role in sending and receiving data. The existence of the smart objects also demands a possession of minimum communication and computation capability. This includes the ability to be discovered, accepting the incoming message and replying to them, discovering services and performing the tasks related to network management. All the above stated functions require a memory range from a few Kilobytes to Gigabytes as per the need. The memory helps in the storage of data. This data is then used for analysis, computations and decision making. 4. Processing unit: A smart object has a processing unit for gathering, processing and analyzing the data acquired from the sensors. The computations done call for control signals that prompt the actuator according to the need. Various functions of the smart object are controlled by the processing unit like communication and power system. The type of processing unit to be used can vary according to the needs and kind of processing to be used by the applications. Microcontroller is the most widely used processing unit due to its small size, simplicity of programming, less power consumption, flexibility and low cost. 5 . Communication unit: All the smart objects need to communicate to each other for the sharing of information. This communication may take place between two or more smart objects or to the outside world through network. The communication is only possible if the smart object has a unit meant for communicating over a wire or wirelessly. The wireless connectivity is preferred more over the wired connectivity for various reasons like cost, ease of deployment and limitations associated with the infrastructure. The communicating unit follows a set of rules, generally termed as protocols to share the information across the network. 6 . Power source: There are various components in a smart object. These components need power to operate. The different sources from which power can be attained are batteries, solar power, wind power, main supply etc. Communicating unit of a smart object accounts for the maximum power consumption. The requirements of power consumption vary greatly according to the scenarios like the area of deployment, switching between the active and sleep mode, accessibility, the power source being used (battery, solar or wind etc), area of application, criticality of the information etc. The different combination of scenarios or conditions calls for different selection of power source.

1.4.2  Trends in Smart Object Smart objects show vast variability in their basic characteristics like technical requirements, functions, conditions of deployment etc. Today we can see and infer the macro trends in smart objects taking the world towards a planned smart future. Some major generalization and trends impacting IoT are:

12

Zaheeruddin and H. Gupta

1. Decrease in size: The size of the sensors are decreasing day by day, The process of miniaturization has excelled so much that in some cases the sensors are not visible from the naked eyes. A trend can be seen in the decrease in the size of the sensors. In general, the smart objects are embedded into everyday objects. The smaller size of the smart objects makes the process of embedding easier. 2. Increase in the processing power: With the passage of time the potential of processors has risen continuously and the size has reduced to a large extent. This advancement in the smart objects have made them more complex and connected. 3. Declination in the power consumed: A smart object comprises of different parts. These parts continuously consume power. The sensors may be active or passive. These days sensors are turning out to be completely passive having no requirement of the external power supply. On the contrary the battery powered sensors are also being designed in such a manner that they are able to last for many years without any need of replacement. 4. Improvement in the communication capabilities: A great improvement can be seen in the speed and range of communication provided by the wireless technologies. With the growth in the use of IoT network, more and more specialized communication protocols are being developed to support various applications and environments. The stated trends in smart object provide ground for the development of sophisticated devices that are capable of performing complex task with increased efficiency. These trends have resulted due to improved inter-object and inter-system communication. The actual power of IoT is visible only when the smart objects are allied collectively in the network of sensors and actuators.

1.5  Features and Challenges of IoT The term IoT relates to the increasing number of smart and linked entities that emphasizes on the innovative opportunities associated with them. Every entity is possessed with some attributes that makes it unique and different from others.The changing nature of the objects is the fundamental property that makes smart object different from other objects. The term ‘SMART’ in smart objects means that these objects are active, can work autonomously, can form a network, are reconfigurable and have a local control of the resources such as energy and data storage. Some major features and challenges associated with IoT are described below: 1. Heterogeneity: In IoT, various types of devices interact with each other to form a network. These devices are heterogeneous in terms of size, shape and functionality (Miorandi et al. 2012). They may include small sensors embedded in a pen to a huge gigantic machine installed in a factory. These devices have diverse capabilities from the view point of communication and computations. Mutual functioning and interaction of such devices in IoT is the most crucial barrier to be dealt with. It involves the interaction between IoT devices, gateway and cloud

1  Foundation of IoT: An Overview

13

(Alam 2012). The different data format generated also needs to be handled aptly (Malhotra et al. 2017). Therefore, the architectural models and protocols assisting the functioning of IoT should be managed efficiently to handle this heterogeneity. 2 . Scalability: IoT is a huge network with an enormous amount of objects being connected to a global information infrastructure. The number of connected devices has increased many folds giving rise to the scalability issue in the IoT.  For a proper communication to take place, a unique sender and receiver have to be recognized along with the identification of an appropriate path or channel. In order to identify the unique sender and receiver, a unique global identity should be provided to them. The IPv4 addressing (32-bit addressing) scheme used till date offers only 4,294,967,296 unique addresses. This count of addresses is not adequate for the future, where the count of IoT devices will outnumber the world’s population (Gusmeroli et  al. 2009) as given below in Table 1.1. The increase in the number of devices will also deteriorate the quality of communication due to an increased number of interconnections established amidst a large number of entities. The increase in the number of devices will also give rise to huge data which will demand for authentic sources to manage the data and information efficiently (Ali et al. 2019a). 3. Ubiquitous exchange of data by means of proximity wireless technologies: In the modern and imminent future, people are and will be addicted to instant access to data. Today’s generation that have been termed as the always-online generation (AO Generation) will exhibit an eagerness for instantaneous gratification and quick fixes. In order to calm down this anxiety for access to anything, IoT will make the most use of wireless technologies. The wireless technology will enable the isolated, living and non-living objects to be networked together. This feature of wireless technology will provide immediate access to nearly the entirety of human knowledge, and incredible opportunities to connect, create and collaborate (Gubbi et al. 2013). The basic necessity for the wireless communication to take place is the availability of spectrum. For a wireless communication to take place, a wireless adapter is required. This adapter converts the data into radio signals. These radio signals are then transmitted with the help of an antenna and received by the wireless receiver. The transmission and receiving of data is only possible when the radio waves are able to travel form transmitter to receptor. The wireless spectrum, therefore works as a medium for the radio signals to travel. Thus, we can say that the wireless spectrum plays the role of oxygen in-­ Table 1.1  Categorization of the installed base units (Millions of units) (Tung 2017) Category Consumer Business: cross-industry Business: vertical specific Grand Total

2016 3963.0 1102.1 1316.6 6381.7

2017 5244.3 1501.0 1635.4 8380.7

2018 7036.3 2132.6 2027.7 11,196.6

2020 12,863.0 4381.4 3171.0 20,415.4

14

4.

5.

6.

7.

Zaheeruddin and H. Gupta

order to keep the network alive and provides it the capability to work at its full potential. The issue that arises with IoT is the availability of an appropriate spectrum. The requisite of the spectrum works as a as an obligation for the acceptance of dynamic/cognitive radio system. Solutions for optimizing the energy consumption: The entities involved in the setting up the network for IoT vary from a small sensor in a handheld device to big sensors in factories. Some sensors work only when required and some may switch between active and sleep mode as required by the application. Many sensors may have the external power supply to work but many of them will be operated on the batteries. In the case of external power supply, communication and computation is a not a hindrance. But in the case, when sensors are battery operated, communication and computation prove to be an overhead. The constraints that are imposed by the operations of the battery are relieved to some extent by the technique of energy harvesting, but still since the energy will constantly be an inadequate source, it needs to be handled with care (Ali et al. 2019b). There arises need in IoT to devise solutions that tend to optimize energy usage (even at the expenses of performance) will become more and more attractive. Capability of being tracked and localized: The term Radio Frequency Identification commonly known as RFID is extensively used in IoT. All the entities that are a part of IoT network need to be identified. This boon of being identified helps them to be tracked with the aid of applications and RFID tags embedded into the objects. RFID is a short range wireless communication technique that uses electromagnetic field for the detection and identification of the RFID tags attached to objects. These tags are then tracked and interrogated for detecting the location of IoT entities in the physical realm (Gubbi et al. 2013). Capability of being self-organized: IoT is merely a network that is a mixed bag of static network and ad hoc network. The major difference that exists between both the networks is in terms of the stability of position of entities. In static network, the location of the physical existence of the object does not change frequently leading to stability in the network. On the other hand, the physical position of the entities involved in IoT often keeps changing, thus carving out the path for complexity. Here comes in the call for a self organizing network that is capable of being planned, organized and maintained on its own. The IoT network is an ad hoc network with intelligence distributed throughout the network. The objects involved in IoT are termed as smart object whose position and functions keep on changing. They are able to work autonomously by reacting to a varied range of different situation with minimum human intervention. As per the request of the user and the need of scenario, smart nodes in IoT arrange themselves autonomously into a short lived ad hoc network. Using this transient network smart nodes gain the capability to share data and perform tasks in coordination with each other. Abilities to execute service and device discovery, auto tuning to protocol behavior and adaptation to the current context without an external trigger are also included in the scenario discussed above. Semantic interoperability and data administration: The increase in the number of objects being connected to IoT has given rise to huge bulk of data. This

1  Foundation of IoT: An Overview

15

massive quantity of data is exchanged and analyzed at infrequent short intervals. This collection of data is unstructured and lacks a standardized format, semantic description of the related content (Alam and Ara 2013; Shakil and Alam 2013). In order to make IoT a powerful network, this raw and unorganized data needs to be converted into useful information by eradicating all the above issues. This will enforce interoperability among different applications. 8 . Privacy conservation and enhanced security: The rationale behind IoT is to provide connectivity between devices and the Internet. Since more devices are being connected potential risks of data breach, malware ingestion, digital burglary etc. gains momentum. The customers and enterprises are quiet apprehensive about security owing to occurrence of various attacks. The surveys and reports regarding the IoT security provide a clear status regarding the threat to the end user data. Due to the basic engrossment into the physical domain, the technology of IoT needs to conserve privacy and be secure by design (Alam et al. 2015). Therefore, the architecture designs, protocols and other methods for IoT solutions should consider security as a key facet to be taken into account. Breach of security in IoT is the sole reason which will setback the users from accepting and using the technology. Hence, security feature should be well placed in order to support wide adoption of IoT (Gandotra et al. 2017; Maple 2017). The issues and remedies associated with the implementation of IoT has been represented in a simplified tabular form in Table 1.2.

Table 1.2  IoT key issues and requisites needed S.No. Challenge 1 Heterogeneity

2 3

4 5 6

7 8

Requisite Special concern should be shown towards the architectural models and protocols in order to support various devices and data generated by them. Scalability A huge address space should be provided so that unique identification of each entity is possible Ubiquitous exchange of data by An appropriate spectrum should made available for ubiquitous data. means of proximity wireless technologies Solutions for optimizing the Designing of IoT devices and solutions that minimize energy consumption the energy usage. Capability of being tracked and Employment of RFID tags, DSRC or any other short localized range communication The node should be able to execute service, device Capability of being self-organized discovery, auto tuning to protocol behavior and adaptation to the current context without an external trigger. Semantic interoperability and Quick conversion of raw and unorganized data from data administration various sources into useful information. Privacy conservation and The architecture designs, protocols and other methods enhanced security for IoT solutions should consider security as the prime feature.

16

Zaheeruddin and H. Gupta

1.6  IoT as an Industrial Commodity Although the concept of IoT is not new, its perception and handling varies from industry to industry. The efficiency of every company is analyzed by their new products in the market. In order to develop the product, projects are undertaken by the industries. In different phases of the project, knowledge, skills, tools and techniques are applied to accomplish the goals in well defined tenure. In case of an IoT project, the completion of such project comprises of effective handling of data and keeping a tab on the scope of high complex systems. The complexity of system is expressed in terms of number of things, velocity of the generation of data, volume of the data generated, geography, diversified manufacturers and analytics. It is observed from the literature survey that no two implementations of the IoT project are same. The reason for this discrepancy lies in different architectures of IoT that are used by the industries. The basic five steps required in any IoT project are shown in Fig. 1.6. 1. Linkage of Objects: The first essence of an IoT system is to connect to different objects such as devices, machines or sensors. On the establishment of connection, the generated data can be collected. The generation of data can vary from application to application. Data may be generated at every small change or event, or periodically. The tenure in which the data is collected will influence the volume of data gathered. 2. Collection of data: The second essence in the project is to capture the generated data. The data collected in the previous stage is required to be sent to the next stage comprising of centralized data centre or cloud. Immediate or predictive

Alter the data in accordance to the target.

Collect the data generated from the devices.

Move the data to the required loca on.

Fig. 1.6  Steps in an IoT project

Meet the consumer's need by delivering the data.

Provide reliability and security to aain superior perfromance, in spite of immense volume of data.

1  Foundation of IoT: An Overview

17

analysis of data can be carried out at the cloud. If required, data may be ported to the main application for further analysis. 3. Transformation of data: As the data is in progression in the third stage, normalization and modification is performed on the data. This modification is done to customize the data as per the application. Various operations that take place in this stage are buffering of data, filtering of data, aggregation and compression of the data. The reduced and useful data is now sent to the next stage of delivery. 4 . Delivery of data: The reduced and transformed data is ready for delivery at this stage. Delivery of data is made to the target end points, applications and analytics platform. The prime requisite for delivery is to eradicate any loss of data. Any three mechanisms namely request-reply, message based or publish-subscribe can be used for delivering the data. Transmission Control Protocol (TCP), Hyper Text Transfer Protocol (HTTP) or Web Sockets can be used as transport interfaces by the message routers. The transport interfaces can provide messaging infrastructure for communication. 5 . Management of data: The data obtained from the above said stages can now be consumed by the companies. This transformed and reduced data can help the companies to increase their efficiency, reduce risk, predict services and support that they can provide. The data across the IoT system is needed to be controlled and synchronized reliably. This would incorporate managing the network, connectivity of the devices, application provisioning and automation (Khan et  al. 2015). A huge volume of data will flow amongst the devices thus having the need for high security. Therefore every enterprise would need to adhere to the strict rules and regulations in this regard.

1.7  Applications A lot of potential is hidden in the network of IoT which makes the development of numerous applications possible. Many applications can be developed and deployed on the basis of IoT. As the saying goes ‘Rome was not built in a day’, similarly advancements take time. It is step by step process in which development plans can be easily made but the deployment of it takes time (Agarwal and Alam 2018). In this section many applications of IoT have been highlighted. Some common and significant applications of the IoT are briefly discussed below: 1. Aerospace and aviation industry: Counterfeit products and elements can easily be identified with the help of IoT. This can improvise the safety and security of products as well as services. For instance, the industry of aviation is susceptible to the issue of the unapproved part of aircraft that is commonly known as Suspected Unapproved Parts (SUP) (Sletten 2000). There is no assurance of the fact that SUP copes up to the necessities of a standard aircraft part and may lack behind in confirming the stringent limitations of quality of the aviation industry. Thus, security standards with respect to the aircraft are seriously violated

18

Zaheeruddin and H. Gupta

by the SUPs. The authentication and security of the aircraft parts can be easily be infringed by forging the documents. Solution to this problem is to introduce electronic pedigrees. These electronic pedigrees will work for a limited group of aircraft parts. In order to improve the safety and operational reliability of aircrafts, the aircraft parts in their lifecycle, will make a record of their origin and critical security related events. These pedigrees will then be stored and secured within RFID tags and a decentralized database. The RFID tags will be attached to the parts. Whenever a new aircraft part has to be installed in the aircraft authentication can easily be done by verifying the digital signatures and comparing the RFID tags pedigree with the database. These steps can upgrade the safety and operational reliability of the aircrafts. 2. Automotive industry: A lot of up gradation can be seen in the automotive industry as generally all the means of transport are getting equipped with sensors, actuators etc. The installed sensors and actuators are known for their increased processing power (Szmelter 2017). Smart things are used for building and implementing the applications in the automobiles. The applications may monitor and report various parameters of the automobiles like proximity of the other vehicles, pressure in tyres etc. The RFID technology is being used to rationalize the production of vehicles, to improvise logistics, to enhance the quality control and improve customer services. The real time data in the manufacturing process and maintenance operations is provided by the RFID technology. This provides effective assistance in managing the product and the automobiles. In order to achieve high bit rate and reduction in the interference with the other equipment, Dedicated Short Range Communication (DSRC) is employed. The applications of the Intelligent Transportation System (ITS) which include traffic management and safety services of vehicle can fully be incorporated in the IoT infrastructure by using communications like Vehicle-to-­ vehicle (V2V) and vehicle-to-infrastructure (V2I) communication. 3. Telecommunications industry: The alliance between dissimilar technologies can be made possible with the help of IoT. This will work as the foundation for creating new services. An illustrative scenario for the above alliance is the mutual use of GSM, Bluetooth, Near Field Communication (NFC), Wireless Local Area Network (WLAN), Global Positioning System (GPS), multi-hop networks and sensor network with Subscriber Identity Module (SIM)-card technology (Deshpande et  al. 2017). The different technologies stated above have significant features which help in the implementation of the services. In these services or applications the SIM-card is attached to the mobile phone. The mobile phone has the reader (i.e. tag) and different applications share the SIM-­card. A secure and simple communication among objects can be enabled by the use of NFC. This is achieved by placing the objects in the vicinity of each other. In this scenario the mobile phone can read all the data by imitating as a NFC-­reader. This data is then transmitted to a central server. The SIM-card in the mobile plays an important role by storing the authentication credentials like ID information, ticket numbers etc. and NFC data. For specialized purposes the robustness of the communication channel and networks can be

1  Foundation of IoT: An Overview

19

increased by facilitating peer-to-peer communication. In order to achieve it the ‘things’ can connect to form a network. An ad-hoc peer-to-peer network can also be formed to handle the situation of disaster or a condition of telecommunication infrastructure failures. 4. Medical and healthcare industry: There are many applications in the healthcare sector that can be framed by using the features of IoT.  Monitoring the medical parameters and tracking the delivery of drugs can easily be done by using cell phones with RFID-sensor capabilities. Some of the advantages that can be achieved by using the said feature are: (a) monitoring of diseases can be made easy, (b) ad-hoc diagnosis can be done, and (c) instant medical aid can be provided in case of accidents. Health records can be saved and secured by using implantable and wireless devices. These health records can be used to save patient’s life and special treatment can be given to people in emergency situation especially those suffering from heart disease, cancer, diabetes, stroke, cognitive impairments, Alzheimer’s etc. Guided action on a body can be taken by introducing biodegradable chips into the human body (Machorro-Cano et al. 2017). Muscular stimuli can be delivered to paraplegic persons for restoring the movement functions. This can be attained by implanting a smart thing-­ controlled electrical simulation system. 5. Pharmaceutical industry: Safety and security of the pharmaceutical products are the prime requirements that need to be fulfilled for the effective use of the drug. Using the technique of IoT smart labels can be attached to the drugs. These smart labels enable the tracking and monitoring of the drug in the supply chain management. The status of the drug can be monitored providing many potential benefits. For instance, many products of pharmacy like vaccines and some drugs are required to be stored at a cool temperature. An appropriate maintenance of the cool chain can be monitored with the help of IoT. In case, if the required cool temperature is not maintained during storage or transportation, then the product can be discarded. The developing countries are affected by counterfeiting of the drugs. Drug tracking and e-pedigrees can be used in order to eradicate fraudsters and detect the counterfeit products (Fantana et al. 2013). 6. Retail, logistics and supply chain management: The operations of retailing and supply chain management (SCM) can benefit from IoT. Many applications can be optimized by the retailer by implanting RFID chips to the products and using smart ledges to follow up the availability of items in real time. For instance, automatic inspection of goods receipt, tracking the goods that are out of stocks, real time monitoring of the stocks and many such activities can be done by the retailer. It has been found that loss in sales is also detected when the customer does not find the required product in the shelves and returns back without the product. This loss of the retail store can be reduced with the help of IoT.  Moreover, the logistics of the whole supply chain management can be optimized by providing the availability of data from the retail store. The availability of stock and sales data from the retailers will help the manufacturers to produce and dispatch the appropriate amounts of product. This will in turn

20

Zaheeruddin and H. Gupta

avoid the situation of over production or under production. The exchange of RFID data can prove beneficial to several sectors of industry by improvising the supply chain incorporating the logistic processes (Da Xu and He 2014). 7. Environment monitoring: Environmental monitoring and conservation can be considered as one of the booming segment in the market in the coming future. Identifiable wireless devices can be utilized for efficient monitoring of the environment. The monitoring may include scrutinizing of the weather condition, level of humidity, level of pollution etc. 8. Transportation industry: IoT has the capability of providing many solutions for improving the transportation industry. It can provide solutions for automatic toll and fare collection, thus reducing the traffic and waiting queues on road. With the help of IoT, the transportation system can be modified by the deployment of Intelligent Transportation System (ITS). ITS will help in the commuting of people and goods efficiently. It will monitor traffic jams and provide free pathways to emergency vehicles. It can also provide support and improve the security policies across the globe by providing means for the automatic screening of commuters and their baggage’s boarding the cargo system (Skabardonis 2008). 9. Agriculture and breeding: A technology like IoT can be used to regulate the traceability of animals used for agriculture purpose. This can assist the detection of animals in real time especially during the eruption of infectious syndrome. In many countries farmers and shepherds are given subsidies as per the basis of the number of animals like cattle, goat, sheep etc. People can do fraud with the government by misleading them with the count of animals. The detection of such deceptions is a difficult task. Therefore, IoT can be employed to reduce the fraud as it can provide appropriate methods of identifications. Many other applications like conducting survey, controlling and prevention of diseases can easily be done by using varied features of IoT. IoT can be used to accurately identify the different specimens of blood and tissues and provide certification regarding the health status of animals in a region. The concept of IoT will also benefit the farmers as they can have a direct contact with the consumers. The farmers can then make direct deliveries of crop to not only smaller regions but also wide area markets. This will provide new way of supply chain and prove beneficial both for the producers and consumers (Maple 2017). 10. Media, entertainment industry: Online videos and news have become very common now days. An enhancement that IoT technology can show in this area is to deploy multimedia devices at different locations. Using these devices ad hoc news can be gathered on the basis of the location of the users. Financial offers can be provided to the people at these locations for collecting the footages. More information can be congregated by affixing the NFC tags to the posters. The tag readers are connected to a URI address possessing exhaustive details about the poster. 11. Insurance industry: The privacy of an individual is very important but IoT technology is considered as a serious hinder in maintaining it. But in order to avail the monetary benefits people are willing to compromise with their privacy.

1  Foundation of IoT: An Overview

21

Let’s take an example of vehicle insurance which has been equipped with ­electronic recorders with the permission of their owners. The advantage of this recorder is that it can record dangerous driving pattern and communicate this information to the car owner or insurer. Similarly in case of house insurance, in-home sensors can monitor the water and fire damage and inform the insurer. The advantage of this kind of insurance is that the insurer can be informed of the imminent damages and can trigger the best possible economic action. Same kind of insurances can be provided on different assets like machinery, factories etc. The innovation provides cushion for maintenance at cheaper rates before the occurrence of the incident. 1 2. Home Automation: These days’ people rely more on technology to address concerns regarding lifestyle led by them and security of their homes. The advancements in the technology of sensors, actuators and wireless sensor network are the main pillars behind the popularity of converting homes to smart homes. In smart homes, intelligent and automated services are provided to the user by the deployment of sensors at different locations. The smart sensors not only automate the day to day tasks of an individual but also help in energy conservation by switching off the electronic gadgets like fans, lights etc. when not in use. Energy conservation in smart homes is made possible by the use of sensors and the concept of context awareness. Different data like (temperature, light, humidity, fire event, gas etc.) are collected by the heterogeneous sensors and fed to the context aggregator. The aggregator passes on the collected data to the context aware service engine. On the contextual basis, different services are selected and the required task is achieved (Pal et al. 2018). For instance, an increase in the humidity level will switch on the AC automatically. A lot of modification in smart homes can be done as and when required.

1.8  Conclusion The living standard in the modern society has moved to a different level by the emergence of Internet of Things. The Internet of Things has enabled communication amongst smart objects, contradicting the previous definition of interaction that was confined to the interactions amongst human and machines. In the present scenario, the term interaction encompasses communication between ‘anything’ irrespective of location, time and type of communication. This chapter introduces IoT as a combination of Information Technology and Operational Technology. IT supports protected connectivity of the data and gadgets within an organization whereas OT administers devices and processes on physical system. The new paradigm of IoT has shown an immense progress in digitization and has framed an aura where everything including people, process, data and things are connected. In order to assist the global connectivity of things new features have to be added to the existing objects and internet. The ordinary objects are upgraded to the smart object by integrating

22

Zaheeruddin and H. Gupta

communication and computation capability. The characteristics, trends and features of IoT objects have been covered in detail in the chapter. Various types of project can be undertaken by IoT involving different inputs, attributes and solutions. Each project can follow a different architecture as per the requirement. In this chapter, two basic possible architectures namely oneM2M and IoT World Forum Architecture have been illustrated. The perception and handling of IoT can vary from industry to industry. Therefore, any IoT project should progress in a sequence employing and modifying the architecture as per the requirement. Each commodity provided by the industry will pave the path for new applications. Upcoming developments and growth in IoT will also optimize the flow of information thus revolutionizing private and business communications. The various applications of IoT are described in the last section of the chapter. With firm determination it can be said that a lot good can be achieved in the field of IoT if the networking and communication research are carried out together in the laboratories of industries and academic institutions complementing each other.

References Agarwal, P., & Alam, M. (2018). Investigating IoT Middleware Platforms for Smart Application Development. arXiv preprint arXiv:1810.12292, pp. 1–14. Alam, M. (2012). Cloud algebra for handling unstrucured data in cloud database management system. International Journal on Cloud Computing: Services and Architecture, 2(6), 35–42. Alam, M., & Ara, K. (2013). A decision matrix and monitoring based framework for infrastructure performance enhancement in a cloud based environment. In International conference, pp. 174–180. Alam, B., Doja, M.  N., Alam, M., & Mongia, S. (2013). 5-layered architecture of cloud database management system 5-layered architecture of cloud database management system. AASRI Procedia, 5(January 2015), 194–199. Alhafidh, B. M. H., & Allen, W. (2016). Design and simulation of a smart home managed by an intelligent self-design and simulation of a smart home managed by an intelligent self-adaptive system. International Journal of Engineering Research and Applications, 6(8), 64–90. Ali, S.  A., Khan, S., & Alam, M. (2019a). Resource-Aware Min-Min (RAMM) algorithm for resource allocation in cloud computing environment. International Journal of Recent Technology and Engineering, 8(3), 1863–1870. Ali, S. A., Member, S., Affan, M., & Alam, M. (2019b). A study of efficient energy management techniques for cloud computing environment. In International conference on cloud computing, data science & engineering, pp. 13–18. Anjanappa, M., Datta, K., & Song, T. (2002). Introduction to sensors and actuators. In R. Bishop (Ed.), The mechatronics handbook (pp. 1–15). Boca Raton: CRC. Bello, O., & Zeadally, S. (2015). Intelligent device-to-device communication in the Internet of Things. IEEE Systems Journal, 10(3), 1–11. Bohn, J., Coroam, V., Langheinrich, M., Mattern, F., & Rohs, M. (2003). Disappearing computers everywhere – Living in a world of smart everyday objects 1. In New media, technology and everyday life in Europe (pp. 1–20). Chemudupati, A., Kaulen, S., Mertens, M., & Zimmermann, S. (2012, November). The convergence of IT and operational technology. Atos Scientific Community (pp. 3–16). Cohen-Almagor, R. (2011). Internet history. International Journal of Technoethics, 2(2), 45–64.

1  Foundation of IoT: An Overview

23

Deshpande, P., Damkonde, A., & Chavan, V. (2017). The Internet of Things: Vision, architecture and applications. International Journal of Computers and Applications, 178(2), 1–14. El Jai, A., & Pritchard, A. J. (2007). Sensors and actuators in distributed systems. International Journal of Control, 46(4), 1139–1153. Fantana, N. L., et al. (2013). IoT applications – Value creation for industry. In Internet of Things: Converging technologies for smart environments and integrated ecosystems (pp.  153–206). River Publishers. Gandotra, P., Kumar Jha, R., & Jain, S. (2017). A survey on device-to-device (D2D) communication: Architecture and security issues. Journal of Network and Computer Applications, 78(July 2016), 9–29. Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013). Internet of Things (IoT): A vision, architectural elements, and future directions. Future Generation Computer Systems, 29(7), 1645–1660. Gusmeroli, K., Haller, S., Harrison, M., Kalaboukas, K., Tomasella, M., Vermesan, O., & Wouters, K. (2009, April). Vision and challenges for realizing the internet of things. In Cluster of European research projects on the Internet of Things (Vol. 1, pp. 44–58). Kallmann, M., & Thalmann, D. (1999). Direct 3D interaction with smart objects. In ACM symposium on virtual reality software and technology (pp. 124–130). Kevin, A. (2010). That’ Internet of Things’ thing. RFiD Journal, 22, 97–114. Khan, I., Naqvi, S. K., & Alam, M. (2015). Data model for big data in cloud environment. In 2nd IEEE international conference on computing for sustainable global development, pp. 582–585. Lannacci, J. (2018). Internet of things (IoT); internet of everything (IoE); tactile internet; 5G – A (not so evanescent) unifying vision empowered by EH-MEMS (energy harvesting MEMS) and RF-MEMS (radio frequency MEMS). Sensors and Actuators A: Physical, 271, 187–198. Leiner, B.  M., et  al. (2009). A brief history of the internet. Computer Communication Review, 39(5), 22–31. Machorro-Cano, I., Alor-Hernandez, G., Cruz-Ramos, N.  A., Sanchez-Ramirez, C., & Segura-­ Ozuna, M.  G. (2017). A brief review of IoT platforms and applications in industry. In New perspectives on applied industrial tools and techniques (pp. 293–324). Malhotra, S., Doja, M. N., Alam, B., & Alam, M. (2017). E-GENMR: Enhanced generalized query processing using double hashing technique through MapReduce in cloud database management system. Journal of Computer Science, 13(7), 234–246. Maple, C. (2017). Security and privacy in the internet of things. Journal of Cyber Policy, 2(2), 155–184. Miorandi, D., Sicari, S., De Pellegrini, F., & Chlamtac, I. (2012). Internet of Things: Vision, applications and research challenges. Ad Hoc Networks, 10(7), 1497–1516. Mühlhäuser, M., & Gurevych, I. (2010). Introduction to ubiquitous computing. In J.  Symonds (Ed.), Ubiquitous and pervasive computing: Concepts, methodologies, tools, and applications (pp. 1–19). Pal, D., Funilkul, S., & Charoenkitkarn, N. (2018). Internet-of-Things and smart homes for elderly healthcare : An end user perspective. IEEE Access, 6, 10483–10496. Perera, C., Liu, C. H., Jayawardena, S., & Chen, M. (2014). A survey on Internet of Things from industrial market perspective. IEEE Access, 2, 1660–1679. Pinelis, M. (2017). Sensors and electronics for the cockpit of the future. Retrieved May 21, 2019, from https://www.smart-mobility-hub.com/sensors-and-electronics-for-the-cockpit-of-the-future Sadiku, M.  N. O., Tembely, M., & Musa, S.  M. (2018). Internet of vehicles: An introduction. International Journal of Advanced Research in Computer Science and Software Engineering, 8(1), 11. Shakil, K. A., & Alam, M. (2013). Data management in cloud based environment using k- median clustering technique. In 4th international IT summit confluence 2013  – The next generation information technology summit (pp. 8–13).

24

Zaheeruddin and H. Gupta

Singh, D., Tripathi, G., & Jara, A. J. (2014, March). A survey of Internet-of-Things: Future vision, architecture, challenges and services. 2014 IEEE world forum internet things, WF-IoT 2014, pp. 287–292. Skabardonis, N. G. A. (2008). Real-time monitoring and control on signalized arterials. Journal of Intelligent Transportation Systems, 12(2), 64–74. Sletten, S. J. (2000). Suspected unapproved parts in the aviation industry: Consideration of system safety and control. The Journal of Aviation/Aerospace Education & Research, 9(3), 11–16. Szmelter, A. (2017). The concepts of connected car and internet of cars and their impact on future mobility. Information Systems Management, 6(3), 234–245. Tung, L. (2017). IoT devices will outnumber the world’s population this year for the first time. [Online]. Available: https://www.zdnet.com/article/iot-devices-will-outnumber-the-worldspopulation-this-year-for-the-first-time/. Accessed 20 Mar 2019. Xu, L. D., He, W., & Li, S. (2014). Internet of things in industries: A survey. IEEE Transactions on Industrial Informatics, 10(4), 2233–2243. Yassein, M. B., Aljawarneh, S., & Masa’deh, E. (2017). A new elastic trickle timer algorithm for Internet of Things. Journal of Network and Computer Applications, 89, 38–47.

Chapter 2

Cloud Computing for IoT Himani Tyagi and Rajendra Kumar

Abstract  The Internet of Things (IoT) combined with cloud computing is an extensively stimulating technologies existing today. IoT is all about sensors, actuator, networks and widely distributed smart devices with limited storage and processing capability with prevalent security and privacy issues. Due to the communication between interconnected devices, the enormous amount of data generated in IoT often referred as Big Data that brings heaps of strain on the internet infrastructure. This has made organization to look for a solution or alternative to reduce this load and introduce cloud computing to solve this problem by providing on demand and virtual services like unlimited storage and processing power. These both technologies are inseparable and work in integration towards increasing the efficiency of every day task. Continuously cloud systems are evolving to provide great support to the Internet of Things (IoT). IoT produces continuous or streaming data and cloud computing on another hand gives meaning and provides path to this data. In addition to this, by providing remote storage and access to data, cloud allows developers to implement projects with no delays. Also, taking advantages of this, many cloud providers provide pay-as-you use strategy and charge users for the services used. This chapter discusses the cloud based support to IoT, applications offered by the paradigm, platforms available, challenges faced by integration. Keywords  Internet of Things · communication · Internet · Big Data · Cloud

2.1  Introduction The current development towards cloud-based infrastructures for Internet of things (IoT) ecosystem is providing a relationship between interconnected devices, data generated from these devices (embedded with sensors, actuators, RFID tags) and the way cloud infrastructure is maintained. In the first phase, The IoT frameworks offer implementation ranging from processing, collecting and deploying up to H. Tyagi · R. Kumar (*) Jamia Millia Islamia, New Delhi, India e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_2

25

26

H. Tyagi and R. Kumar

s­ ervice delivery. The next phase is to deploy horizontal IoT frameworks that support on demand access to the deployed framework. Lastly, the cloud infrastructures must support interactions in order to share data and manage operations (Alam 2012a). The Internet of Things brings the real and virtual world tighter than ever before (Atlam et al. 2017). This chapter discusses the cloud based support to IoT and existing cloud based IoT platforms. A study by Cisco showed that the number of networked devices in 2012 was estimated to be 9 billion and is expected to reach 24 billion by 2020. Thus, proving IoT as trending and the main source of big data (Doukas and Maglogiannis 2011). This big data has real values to be analyzed using cloud computing platforms. It generates a huge amount of unstructured and structured data having three characteristics Volume (refers to data size), velocity (refers to data generation frequency), variety (refers to data types) (3V) (Yu Liu et al. 2015). IoT applications are built on the principle of mobility and distributed networks. The deployment requires reliable communication network. Cloud systems play a crucial role in providing utility driven integrated solution for cloud based enterprise services. Despite having many cloud computing models (explained in proceeding section) and infrastructure it is very difficult to manage and control IoT-Cloud based environment (Alam 2012a). Therefore, cloud based IoT applications should be build to make the most out of their combination. Cloud computing is a popular service that offers numerous benefits to IoT, by allowing users to perform normal computation tasks using services hosted on the Internet (Botta et al. 2016). An employee may need to finish a major project that must be submitted to a manager but they encounter with memory and space constraints that can be minimized if an application is hosted on the internet. The employee can use cloud computing services because data is managed remotely by a server. IoT and cloud computing are ‘tightly coupled”. The combination of these two will enable monitoring services and processing of sensory data generated by devices embedded through sensors. IoT can be stated as “a world-wide network of interconnected object that are uniquely addressable (through IP addresses), with the point of convergence as internet. It is totally based on intelligent and self-configuring nodes that are capable of computing, interconnected in a global network infrastructure, responsible to collect, send and receive data. Moreover, Complexity of this network not only includes mobile devices but objects like food, clothing, paper, furniture, watches, cars etc. (Botta et al. 2016) the data generated by these devices put a lot of strain on internet infrastructure. So, this pushes companies to find a solution to minimize this pressure and solve the problem of transferring data. Accordingly, Cloud computing has become one of the major paradigm of IoT and their association is sometimes referred as cloud-IoT paradigm (Botta et al. 2016; Yu Liu et al. 2015). This combination has led to success of the IoT.  IoT projects have additional complexity as compared to other cloud centric technology applications likely 1 . Different gateway network requirement. 2. Diverse operating systems. 3. Diverse hardware requirements.

# of publications

2  Cloud Computing for IoT 100000 10000 1000 100 10 1 2008

27

2009

2010

Year

2011

IoT (content) Cloud (content) IoT (title) Cloud (title)

2012

2013

Fig. 2.1  Popularity metrics of IoT with respect to cloud (Botta et al. 2016)

4 . Diversity of devices involved. 5. Difficult to manage data generating devices and power constrained small sensors (Stergiou et al. 2016). With the advanced technology IoT, anything (person, bulb, food, any physical object) which is beyond imagination can become part of internet and generates useful data. Moreover, data generated needs to be processed in order to create more valuable services (Stergiou et al. 2016). For this purpose, integrating IoT with cloud generates new avenues in information and technology field. Interest and current trend in cloud computing and IoT by content and title for IoT and cloud separately is shown in Fig. 2.1 (Botta et al. 2016) from the year 2008 to 2013.It depicts the increment in interest in both the technology within a time period of 5 years.

2.2  Basic Concepts What if umbrella could sense the weather and provide advises to the user, or if some wearable device could monitor patients’ health or if a car could have some predictive analysis about the upcoming service schedules to avoid certain prompt failures beforehand. IoT and internet-connected-cloud platform as a service (PaaS) can make the above said situation true in real life. The large magnitude of data generated from large number of devices especially from wireless sensor networks, health devices, machine components in industries can be collected, stored, accessed using cloud services. Cloud management is a complex task (Botta et al. 2016).

2.2.1  Cloud Computing Microsoft Azure defined Cloud Computing as “the delivery of computing services-­ including servers, databases, networking, software, analytics, storage and intelligence over the internet (“the cloud”) to offer faster, innovation, flexible resources (Kim et  al. 2016). Cloud computing has changed the way businesses view of IT

28

H. Tyagi and R. Kumar

resources (Atlam et al. 2017; Kim et al. 2016). It is required that cloud computing should reduce computation and processing cost and time. The organizations are adopting Cloud Computing. Here are some common reasons organizations are turning to cloud computing services: 1. Cost Cloud computing has eliminated the capital expenses of buying hardware, software datacentres, servers, electricity for power and cooling. Additionally, it has removed the cost of IT experts for managing the infrastructure. 2. Performance Cloud should support and provide high performance to large groups of diverse users (Alam 2012a). Interoperability and scalability between heterogeneous and distributed internet connected objects (actuators, sensors, etc) is always a challenging task. 3. Speed High speed services offered by cloud computing and vast amounts of computing resources can be provisioned with just a few mouse clicks for IoT based applications makes it adapted by organizations. 4. Security The security and privacy of user information are the most challenging task identified in IoT framework (Atlam et al. 2017). Cloud provides policies and technologies that strengthen overall security (Kim et al. 2016).

2.2.2  Importance of Cloud Computing for IoT Though IoT lacks scalability, interoperability, reliability, scalability, flexibility, availability and security.Cloud has provided these all. It has proved to provide the flow of data, processing and collection while maintaining low cost making data driven decisions and algorithms for predictions possible. The motivation behind combining these two technologies depends on three factors communication, storage and computation. Their challenges and solutions provided are discussed in Fig. 2.2 (Table 2.1). There are some reasons listed below for considering cloud essential for the success of IoT (Fig. 2.3). 1. Provides remote processing power:- cloud empowers IoT to get beyond home appliances such as refrigerators, air conditioners, coffee makers etc. with the advancement in internet technologies like 4G to higher internet speed, the cloud will allow developers to offload fast computing process. 2. Provides security and privacy:-Security in terms of Cloud Computing refers to all the policies to controls and protects data, applications and associated infrastructure. IoT skills are incomplete without security features. Often, the devices bought by users are equipped with default passwords provided by the manufacturers that are easily approachable by eavesdropper. Sometimes, the users don’t

2  Cloud Computing for IoT

29

Fig. 2.2  Cloud IoT drivers Table 2.1  Companionship between IoT and Cloud IoT[1,3] Source of data (data generated by sensors) Act as a point of convergence Limited storage and computing capabilities.

Cloud computing[1,3] Manager of data(storage, collection, computation) Acts as means of delivering services Unlimited storage and computing capabilities.

Fig. 2.3  Reasons for considering cloud beneficial to IoT

use strong enough password to protect their devise and leads to serious problems. Cloud has made IoT more secure by considering protective, detective and corrective controls.It has enabled users with authentication and encryption protocols and provide strong security measures. Many other methods like biometrics are also used to protect users’ identity.

30

H. Tyagi and R. Kumar

3. Provides massive data storage:-In IoT there are millions of devices involved that are continuously or periodically generating data. The data that is outside the capacity of normal systems to handle. This is an open challenge for IoT and cloud together to handle and store the data. Two different cloud models based on usage: • Service Models (SaaS, PaaS and IaaS) • Deployment Models (Private, Public and Hybrid) 4. Facilitates inter-device communication:- cloud behaves as a mediator when comes with IoT communication. The data generated by devices can be analysed and processed on cloud for making intense decisions. 5. Remove entry barriers to the host providers:- With cloud, most hosting providers can allow their clients a ready-to-roll model, removing entry barriers for them.

2.3  Cloud Based IoT Architecture 1. Perception layer It is responsible for collecting all information from its surroundings (sensors and actuators) and sends it through gateway for further processing. The attacks on sensors are the most common and constant in IoT framework. Here, the parameter values are evaluated by RFID tags, sensors, GPS, RSN other devices (Fig. 2.4). 2. Network layer Based on the distance for communication the protocols are used as Bluetooth (small distance communication), ZIGBEE/LAN (communication within a ­building), WiFi (communication from anywhere in the world).the security con-

Fig. 2.4  Cloud-IoT Architecture[18]

2  Cloud Computing for IoT

31

sideration here are based on data as well as on network. The attacks on network includes WIFI sniffer, malware, eavesdrop, brute force, Man In Middle attack, Man at end attack (MATE) (Vermesan and Friess 2014). Attacks on data includes modification, non-repudiation and provides privacy, confidentiality, availability, integrity as major security considerations. 3. Application layer The advanced layer which is responsible for all classes of business services and intelligent computation, resource allocation and data processing (Rochwerger et al. 2009), where the user can interact with smart devices and control according to their demand. Application layer protocols specially designed for IoT are HTTP (Hyper Text Transfer Protocol) and CoAP (constrained application protocol).

2.3.1  Models The introduction of integrated CloudIoT enables new smart applications [33] based on the extended cloud scope to deal with real life states or scenarios providing way to the birth of things as a service. Some of the new models introduced by this paradigm are given below: 1. CSaaS (Cloud-based-sensing as a Service) (Carnaz and Nogueira 2016):-provides multi-tenancy, virtualization and dynamic provisioning. Making it possible to abstract the data between sensor provider and consumer even sharing the same infrastructure. 2. DBaaS (Database as a Service) (Chen et al. 2010):-allows ubiquitous database management. 3. DaaS (Data as a Service) (Chen et  al. 2010):-allows ubiquitous access to any kind of data. 4. EaaS (Ethernet as a Service) (Chen et al. 2010):-allows ubiquitous layer-2 connectivity to remote devices. 5. SenaaS (Sensor as a Service) (Chen et al. 2010):-allows ubiquitous management of remote sensors. 6. MBaaS (Mobile backend as a Service):[30] provides web and mobile applications to connect to the backend cloud storage. It Provides features like user management, social networking services. 7. VSaaS (Video Surveillance as a Service)[31]: provides all video storage requirements like stored media is secured, scalable, on demand, fault-tolerant and video processing. 8. SaaS (Sensing as a Service) (Chen et al. 2010):-allows ubiquitous access to sensor data.

32

H. Tyagi and R. Kumar

2.4  CloudIoT Applications The integration of cloud with IoT brings many opportunities as well as challenges that are discussed in proceeding section. The CloudIoT paradigm is a great evolution in Information and Communication Technology (ICT). This paradigm has taken many industries including agriculture, healthcare, energy management, transportation, smart city, environment protection that are discussed in next chapters. Cloud has provided the ability to store and stream data over the internet, possible visualization in a web browser, controlling devices from anywhere and anytime in the world (Prati et al. 2013). The usage of cloud in IoT has acted as a catalyst for the development and deployment of scalable applications. The two technologies provide platform to each other for success (https://azure.microsoft.com/en-in/overview/what-is-cloud-computing/). Cloud computing has brought a revolution how technologies can be managed, accessed and delivered (Atlam et al. 2017). The integration has given birth to numerous applications as depicted by Fig. 2.5 and most of the applications affects everyday tasks. Cloud can provide benefits to IoT by extending its scope to deal with real world things in a more dynamic and distributed way. In most cases, Cloud can provide the middle layer between the things and applications, abstracting all the complexity and functionalities necessary to implement the applications (Botta et al. 2016). In this section, we describe the applications offered by CloudIoT paradigm (Fig. 2.6 and Table 2.2).

Fig. 2.5  Services offered by CloudIoT paradigm (Source: https://siliconangle.com/)

2  Cloud Computing for IoT

33

Fig. 2.6 Applications offered by CloudIoT paradigm[1]

Table 2.2  IoT applications description IoT applications Importance of integration of cloud and IoT Health care Collecting patient’s important data (ECG, BP, Heartbeat) via network of system sensors connected to human body, delivering this data to medical centre’s cloud for storage. further processing and sharing medical report data (Botta et al. 2016). It can improve health care services and health care processes by providing cost effective and high quality services. It also aims at providing Ambient Assisted Living (AAL) (Kaur and Alam 2013) means easing the daily lives of old person, person with physical and mental disabilities. Smart city The smart city solutions should cover the notions like, Smart Lighting, Smart traffic management, Smart building, Smart parking, Wi-Fi Internet access & City Surveillance, Solid Waste Management, Smart Metering, Water Quality, water clogging management in cities, etc. It also aims at easing the urban life, protecting environment, encourage sustainable development (Alam 2012b). For the efficient and flawless operation of smart cities, deployment of IoT and cloud computing is most important. It is impossible to offer above mentioned services and visualization, management, processing of enormous amount of vital data that is useful for public or private organization without cloud. (continued)

34

H. Tyagi and R. Kumar

Table 2.2 (continued) IoT applications Importance of integration of cloud and IoT Smart home and Cloud has automated all the common household activities. The association of smart metering computing with physical objects transforms the environment into informative and appliances into intelligent. Smart lighting is responsible for 19% global use of electrical energy and emission of greenhouse gases (Alam 2012b). To emphasis on sustainable development this emission and usage should be reduced. Smart lighting control systems has proved to reduce the light consumption by 45% (Botta et al. 2016). Cloud offers flexible framework for building smart applications making home automation a trivial task. Cloud offers remote control of appliances and services. Cloud offers direct communication of users and sensors. Cloud can satisfy that any digital appliance can be connected to any other. The intelligent video surveillance make it possible to monitor anything at Video surveillance anytime from anywhere. CloudIoT in this context can be used to store, manage and process video contents evolving from IP cameras (Botta et al. 2016; Atlam et al. 2017) and automatically extract knowledge. Because of cloud, Security can now be offered as a service that is managed remotely and freely that no longer require human. VSaaS[31] given cloud model satisfies all the video storage requirements i.e, the media is centrally secured, scalable, fault tolerant and also provides video processing facilities to identify patterns. Automotive and Intelligent transport system (ITS) is offering a great opportunity in the field of Smart mobility transportation system and automobile industry by providing traffic state prediction, notification (Atlam et al. 2017). The integration of Cloud with IoT technologies (WSN and RFID) brings many opportunities (Botta et al. 2016). Vehicular cloud data mining service provides road safety, reducing road congestion, smart parking system, recommending car maintenance system. Vehicle to vehicle communication, vehicle to infrastructure communication is possible [30,32]. Smart energy The integration cloud with IoT has an excellent application in the field of and Smart Grid smart energy and smart grid by providing intelligent energy distribution and consumption (smart meters, smart appliances, renewable energy resources) (Botta et al. 2016; Atlam et al. 2017). Smart Logistics The CloudIoT paradigm in the field of logistics provides automated management of flow of goods from the point of source(origin) to the point of destination in a cost effective and timely manner (Botta et al. 2016). It has made the tracking in transit possible effortlessly (Botta et al. 2016; Atlam et al. 2017). The implementation of Cloud computing can overcome the bottleneck by allowing complex decision-making systems with automated algorithms to retrieve information for further planning and tracking. Environmental The integration offers a great scope in the area of environment monitoring by Monitoring deploying environmental sensors in any area. some applications include water level needed in field, lighting condition, gas concentration in air, natural hazards occurrence like landslides, detecting infrared radiation (Botta et al. 2016; Atlam et al. 2017).

2  Cloud Computing for IoT

35

2.5  Cloud Platforms Available for IoT A number of IoT platform are available for sensor data management either free or open source (ThingSpeak, iCloud, Dropbox) or commercial (AWS, Azure, Kura, SmartThings, google cloud platform).

2.5.1  Commercial IoT Platform 1. Aws IoT platform AWS IoT (https://ap-south-1.console.aws.amazon.com/console/ home?region=ap-south-1), A cloud platform provided by amazon for the internet of things (IoT). This framework provides smart devices to connect easily and securely interact with AWS (Amazon Web Services) cloud. The most exciting feature of AWS is it allows devices to communicate even when they are offline (https://azure.microsoft.com/en-in/solutions/). This framework offers services like for computing (Lambda, EC2), for storage (Amazon S3, storage gateway), database (DynamoDB), machine learning (Amazon DeepLens, Amazon forecast), for Analytics (kinesis), blockchain (Amazon Managed Blockchain) (Ammar et al. 2018) (Fig. 2.7 and Table 2.3). 2. Microsoft Azure Azure is a platform offered by Microsoft that allows end users to interact with IoT devices, receive data, performs aggregation, multi-dimensional analysis and transformation. AWS, as discussed and Azure are two famous names in public

Fig. 2.7  Cloud providers to IoT

36

H. Tyagi and R. Kumar

Table 2.3  AWS Description and Applications Aws IoT components[16] Description Cloud services Kinesis, DynamoDB, Blockchain, Lambda. Smart devices Anything ranging from human being to bulb. Device gateway Supports application protocols like MQTT, Web Socket, HTTP 1.1 Message broker Registry Rule engine

Rules like SQL syntax queries

Language support Cryptography Security

Any language 128 bit AES X.509 certificate (Eclipse Organization 2019), AWS cognito identities SSL/TLS

Communication (SmartThings 2019) SDK language support Device Shadow

C, node.js, Arduino JSON document

Applications Machine learning, computation, blockchain, Database, storage

Provides security, manageability, and scalability Publishing and subscribing messages Tracking and identification To devices Processing, analyzing, Gathering, action on data Supports Security Authentication

Confidentiality and encryption Allow accessing any device Retrieve current state of certain device

cloud computing. It supports wide range of operating system, programming languages. IoT devices interact with azure cloud through a gateway (Eclipse Organization 2019). The data from devices is stored on cloud for further processing and analysis by applying machine learning service offered by azure. The real time monitoring of data is also possible (Table 2.4). 3. SmartThings SmartThings is a IoT cloud platform released by Samsung mainly has applications in smart homes by enabling users to have control over smart devices.it is a platform where developers can implement applications allowing their users to control smart appliances in homes through mobile applications (Alharbi and Aspinall 2018) (Table 2.5). 4. Kura platform It is an extensible open source Java/OSGi (open services gateway initiative) IoT Edge framework. It is an advanced software framework that abstracts the developer from hardware complexity and complexity of networking sub-systems. It is a project that provides platform for building IoT gateway (Jing et al. 2014). It is a smart application container, which provides re0mote management of the gateways and APIs for building your own IoT application (Table 2.6).

2  Cloud Computing for IoT

37

Table 2.4  Azure platform Description and Applications Azure components Cloud services Smart devices

Description Anything ranging from human being to bulb

Field gateway Azure IoT hub

Commands/notification

Identity Registry Security SDK language support Azure access Directory communication

Applications Machine learning, Azure stream analysis

X.509/TLS based handshake and encryption C, Node.js, python, java, .NET

SSL/TLS

Aggregation, storing and forwarding data to IoT hub. Provides bi-directional communication. Provides device Identification and Management Provides authentication Provides device management Provides authorization and access control Encrypt communication and ensures confidentiality and integrity of data.

Table 2.5  SmartThings Description and Applications SmartThings components SmartThings mobile client app SmartThings cloud backend SmartThings Hub

Description Supports Android and iOS Abstraction and intelligence

communication

128-bit AES encryption, Zigbee product Should facilitate connection via WiFi/IP protocol Supports communication protocol Z-wave, WiFi, Zigbee, BLE SSL/TLS

Security protocol

OAuth/OAuth2 protocols

Smart devices Home controller

Applications Smoothly access the smart appliances. Hosting and running smart apps, running virtual software image of physical smart devices. Security enabled Z-wave product Connecting these devices to cloud for further processing Gateway for communication between devices and cloud Encryption communication to ensure confidentiality and integrity Authentication

Table 2.6  Kura Platform Description and Applications Kura components OS support Programming language support Supported protocols Communication Device abstraction layer

Description[18] Only Linux java

Cryptography

Multiple cryptography primitives

MQTT, CoAP SSL/TLS Using OSGi services

Applications Provides remote management For developing smart applications Authenticate communication Secured communication Abstracting hardware complexities

38

H. Tyagi and R. Kumar

2.5.2  Open Source IoT Platform • ThingsBoard:- An open source platform for processing, collection, visualization and management of data produced by sensors communication involved in IoT ecosystem (Table 2.7). • The devices are connected via IoT protocols MQTT, CoAP (constrained application protocol) and HTTP which also supports cloud deployments. ThingsBoard provides scalability, fault-tolerance and performance to the data. It supports both cloud and on premises deployments. You just have to create account and manage your smart devices smartly. There is special provision to raise alarms on incoming telemetry data. Figure 2.8 shows real time monitoring of smart meters IoT system on ThingsBoard platform. Table 2.7  ThingsBoard IoT platform ThingsBoad components IoT rule engine Real time IoT dashboard Security

Description Allows to create complex chains of rules for data processing from your devices Realtime visualization and remote ontrol of data Supports device authentication management, encryption for both HTTP and MQTT

Fig. 2.8  Showing cloud platform for smart grid application

2  Cloud Computing for IoT

39

Table 2.8  Challenges in applications

Application Healthcare system Smart city Smart home and metering Video Surveillance Automotive and smart mobility Smart energy and smart grid Smart logistics Environment monitoring

Challenges Legal and social Security aspects √ √ √ √

Large scale √

Privacy √ √

√ √ √ √

√ √

√ √ √



Reliability √ √ √

Heterogeneity √ √ √

√ √

√ √









2.6  Challenges In the previous sections the benefits of cloud and IoT integration has shown. This can be accurately and confidently said that the integration is a buzz in technological industry which brings tremendous amount of business and job opportunities. But the integration also brings certain serious challenges that needs to be discussed. Based on the applications discussed the challenges can be summed up as shown in the Table 2.8.

2.7  Conclusion This can be accurately said that cloud can accelerate the growth of IoT. However, deploying cloud technology also comes with certain challenges and shortcomings. Cloud computing and the IoT are the next-generation technologies. Cloud offer an effective solution for IoT like service management, implementing applications and services. It can benefit IoT by enabling it to deal with real world things in a more efficient and scalable manner. It can work as a middle layer between things and applications. Both technologies fills gap for each other like cloud provides IoT with unlimited storage capabilities and IoT extend the scope and application of cloud. The integration of these two technologies provides path for various business opportunities and innovative research areas as discussed in the chapter. The applications presented by integrated technology and challenges derived are discussed in the chapter. The security challenges pertaining to CloudIoT paradigm are power and

40

H. Tyagi and R. Kumar

efficiency, security and privacy. Future scope include finding out a way to deal with heterogeneous environment providing scalability and security, environmental sensors for providing better results, identification of definitive solution for addressing things uniquely, extended support to multi networking.

References Alam, M. (2012a) Cloud algebra for cloud database management system. In The second international conference on Computational Science, Engineering and Information Technology (CCSEIT-2012), Coimbatore, India, Proceeding published by ACM, October 26–28, 2012. Alam, M. (2012b, December) Cloud algebra for handling unstructured data in cloud database management system. International Journal on Cloud Computing: Services and Architecture (IJCCSA), 2(6), ISSN: 2231-5853 [Online]; 2231-6663 [Print], https://doi.org/10.5121/ ijccsa.2012.2603, Taiwan. Alharbi, R., & Aspinall, D. (2018). An IoT analysis framework: An investigation of IoT smart cameras’ vulnerabilities. IoT, 2018. Ammar, M., Russello, G., & Crispoa, B. (2018). Internet of things: A survey on the security of IoT frameworks. Journal of information security and Applications, 8–27. Atlam, H.  F., Alenezi, A., Alharthi, A, Walters, R.  J., & Wills, G.  B. (2017). 2017 IEEE international conference on internet of things (iThings) and IEEE green computing and communications (Green Com) and IEEE cyber, physical and social computing (CPSCom) and IEEE smart data (Smart Data), IEEE. pp.  671–675. https://doi.org/10.1109/ iThings-GreenCom-CPSCom-SmartData.2017.105 Botta, A., de Donato, W., Persico, V., & Pescapé, A. (2016). Integration of cloud computing and internet of things: A survey. Future Generation Computer System, 56), 684–700. Carnaz, G., & Nogueira, V. B. (2016). An overview of IoT and healthcare. In S. Abreu & V. B. Nogueira (Eds.), Actas das 6as Jornadas de Informática de Universidade de Évora. Natal: Escola de Ciências e Tecnologia. Chen, P., Freg, C., Hou, T., & Teng, W.-G. (2010, December). Implementing RAID-3 on cloud storage for EMR system. In Proceedings of the 2010 International Computer Symposium (ICS), Taiwan, 16–18 (pp. 850–853). Doukas, C. & Maglogiannis, I. (2011). Managing wearable sensor data through cloud computing. In Third IEEE international conference on cloud computing technology and science, 2011. Eclipse Organization. (2019). Kura framework. http://www.eclipse.org/kura/. Online. Accessed July 2019. Jing, Q., Vasilakos, A., Wan, J., Lu, J., & Qiu, D. (2014). Security of the internet of things: Perspectives and challenges. Wireless Networks, 20, 2481–2501. https://doi.org/10.1007/ s11276-014-0761-7. https://thingsboard.io/, 2014. Kaur, A., & Alam, M. (2013). Role of knowledge engineering in the development of a hybrid knowledge based medical information system for atrial fibrillation. American Journal of Industrial and Business Management, 3(1), 36–41. https://doi.org/10.4236/ajibm.2013.31005. Kim, M., Asthana, M., Bhargava, S., Iyyer, K.  K., Tangadpalliwar, R., & Gao, J. (2016). Developing an on-demand cloud-based sensing-as-a-service system, internet of things. Journal of Computer Networks and Communications. https://doi.org/10.1155/2016/3292783,2016. Article ID 3292783, 17 pages. Prati, A., Vezzani, R., Fornaciari, M., & Cucchiara, R. (2013). Intelligent video surveillance as a service. In Intelligent multimedia surveillance (pp. 1–16). Berlin/Heidelberg: Springer.

2  Cloud Computing for IoT

41

Rochwerger, B., Breitgand, D., Levy, E., Galis, A., Nagin, K., Llorente, I.  M., Montero, R., Wolfsthal, Y., Elmroth, E., Caceres, J., Ben-Yehuda, M., Emmerich, W., & Galan, F. (2009). The RESERVOIR model and architecture for open federated cloud computing. IBM Journal of Research and Development, 53(4). SmartThings. (2019). Smart things documentation. http://docs.smartthings.com/en/latest/. Online. Accessed July 2019. Stergiou, C., Psannis, K., Kim, B.-G., & Gupta, B. B. (2016). Secure integration of internet-of-­ things and cloud computing. Future Generation Computer Systems, 2016, 964–975. Vermesan, O., & Friess, P. (2014). Internet of things from research and innovation to market deployment. Aalborg: River Publishers. Yu Liu, Beibei Dong, Benzhen Guo, Jingjing Yang, & Wei Peng. (2015). Combination of cloud computing and internet of things (IOT) in medical monitoring systems. International Journal of Hybrid Information Technology, 8(12), 367–376.

Chapter 3

Open Service Platforms for IoT Preeti Agarwal and Mansaf Alam

Abstract  With the advent of Internet of Things (IoT), anything on earth with embedded processor, storage, and communication technology can communicate with each other. IoT can interconnect billions and trillions of devices on earth. This is considered as next revolution in the world of internet. It is expected that this revolution will drastically improve quality of daily life, will bring new forms of collaboration, interaction, and activities. IoT is not considered a single technology, but it is an aggregation of various underlying technologies, making the application development task bit challenging. In order to cope up with this challenge, number of vendors are coming up with IoT platforms for application development. IoT platforms provide support for connecting, storing, computing, and analysing data from heterogeneous devices. This chapter, presents a reference architecture for IoT service platforms, outline a set of service and architectural requirements for IoT platform, and review four major IoT platforms (AWS IoT platform, IBM Watson platform, Microsoft Azure IoT Platform, and Google Cloud Platform) from these requirements viewpoint. Further, gaps and issues in present IoT platforms are discussed with future research directions. Keyword  Internet of Things (IoT) · IoT platform reference architecture · IoT platform service requirements · IoT platform architectural requirements

3.1  Introduction The term “Internet of Things” was first coined by Kevin Ashton in 1999 (Ashton 2009). According to IoT concept, every sensor, every device, and every software can be connected to each other. These devices can communicate remotely with each other via. IoT platform. The main idea behind IoT was to provide ubiquitous

P. Agarwal (*) · M. Alam Department of Computer Science, Jamia Millia Islamia, New Delhi, India e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_3

43

44

P. Agarwal and M. Alam

computing with minimum human intervention (Al-Fuqaha et al. 2015; Atzori et al. 2010). The concept gained popularity by embedding processors, storage and ­communication technology in devices, and these technologies converted devices into smart devices. Smart devices are now capable of sensing the environment, storing information, and can also communicate with each other, eliminating human in the loop. Enabling interaction, collaboration and communication among various devices has foster application development in many domains such as healthcare, smart homes, agriculture, traffic, and energy management (Asghari et al. 2019). In order, to provide centralized control over these smart devices, number of vendors came up with different platforms. At present, market is flooded with IoT platforms, and identifying a suitable one for a particular application is a big challenge (Agarwal and Alam 2019). Some of the popular platforms are AWS IoT, IBM Watson, Microsoft Azure, Google cloud IoT. The main objective of this chapter is: • • • •

To provide general IoT reference architecture, Identify key service and architectural requirements for IoT platforms, Review four most popular IoT platforms from requirements view point, Present gaps and challenges to be addressed by future IoT platforms.

The structure of chapter is as follows: Sect. 3.2 gives general IoT reference architecture, Sect. 3.3 provides key service and architectural requirements of the platform. Section 3.4 discusses four case studies of the most popular IoT platforms AWS IoT, Microsoft Azure IoT, Google Cloud IoT, and IBM Watson IoT from requirements viewpoint. Finally, Sect. 3.5 presents challenges and open research directions to be addressed by future IoT platform.

3.2  IoT Reference Architecture IoT is considered next revolution in World Wide Web. To provide Quality of Service (QoS), special consideration need to be taken to while defining its architecture (Fahmideh and Zowghi 2020). Many IoT architectures have been presented in literature, but the most accepted one is four layer architecture consisting of: sensor layer, gateway and network layer, service management layer, and application layer (Yaqoob et al. 2017), as shown in Fig. 3.1. Each layer is described in following subsections.

3.2.1  Sensor Layer The first layer, called as sensor layer or perception layer. It consist of sensors and actuators, that sense the environment, collect information for processing to gain useful insights. Different kind of sensors can be deployed at this layer, like temperature, motion, humidity, sensing events etc. At this layer heterogeneous devices are deployed in plug and play manner. The perception layer digitalizes, creates secure

45

3  Open Service Platforms for IoT

Applicaon Layer

• Provide services required by customer. • Span various sectors such as Smart Home, Smart Healthcare, Smart Transportaon, etc.

Management and Service Layer

•Provide IoT applicaon programmers with

Gateway and Network Layer

• Connect IoT devices with cloud technologies. • Number of communicaon protocols like Zigbee,

Sensor and Connecvity Layer

• Device sense the enviroment. • Transfers event through secure channel.

hardware abstracon. • Provide features for analysing data.

WiFi, Bluetooth, etc. work at this layer.

Fig. 3.1  IoT architectural layers

channel, and transfer data to the next layer. This layer is major source of big data to be processed by next layers.

3.2.2  Gateway and Network Layer The gateway and network layer transfers data created through secure channels by the sensor layer to the above layer. Various wireless technologies such as Zigbee, RFID, Wi-Fi, etc. are used to transmit data. Furthermore, storage and processing of data needs to be addressed at this layer. One of the possible solution for storage is cloud. Multiple cloud based storage solutions for IoT generated big data exist (Khan et al. 2017; Alam and Shakil 2016). Different ways to process both structured and unstructured data exist (Alam 2012). Most of the data management as well as resource management is carried out at this layer. Many algorithms for efficient resource management can be deployed (Ali et al. 2019).

3.2.3  Service Management Layer Management or Middleware layer binds a service with its requestor. This layer provides features that enables the IoT application programmers to work with the sensor objects in a seamless manner, without any concern to underlying hardware. Also, this layer processes received data, make smart decisions, and based on decisions

46

P. Agarwal and M. Alam

deliver the services over the network through protocols. Various analytical solutions can be applied at this layer to provide intelligent decisions.

3.2.4  Application Layer The application layer provides the requested services to its users. For example, the application layer can provide acceleration and heart beat values to the medical care provider for its patient to continuously monitor them. This layer has ability to provide superior services to meet the user’s need. The application layer provide services to many sectors such as smart home, smart healthcare, smart transportation, industrial automation and smart energy management.

3.3  IoT Platform Requirements IoT platform is a responsible for integrating devices on network for different applications through software packages. IoT Platforms are deployed on service management layer of the IoT architecture. The platforms provide users a layer of abstraction, hiding the implementation details. IoT platforms provides an ecosystem upon which different smart applications can be built (Tiwana 2013). Like other software’s, IoT platforms also need to satisfy certain user requirements. These user requirements are divided into service requirements and architectural requirements (Razzaque et  al. 2015; da Cruz et  al. 2018). The service requirements can be functional and non-functional, as shown in Fig. 3.2.

3.3.1  Service Requirements Service requirements can be classified as functional and non-functional. Functional requirements describe the required functionalities of the IoT platform, whereas nonfunctional requirements focus on providing QoS parameters (Razzaque et al. 2015). The various functional requirements are described as follows: • Resource Discovery: The IoT connects heterogeneous devices in dynamic environment. There must be some automated mechanism to publish, discover, and subscribe to resources in centralized manner. In order to carry out this task middleware platform maintains a registry component in which devices register themselves using necessary API’s. After registration the device becomes discoverable to other devices in the network. The registered device can publish metadata about its services. Even other devices in the network can query the device by suitable query mechanism, can surf the required services and resources provided by the

3  Open Service Platforms for IoT Fig. 3.2  IoT platform requirements

47

Funconal Requirements

•Resource Discovery •Resource Management •Data Management •Data Acquision •Data Cleaning •Data Storing •Data Processing •Query Processing •Data Visualizaon •Event Processing •Code Management

NonFunconal Requirements

•Scalability •Timeliness •Availability •Security •Privacy •Ease of deployement

Architectural Requirements

•Programming Abstacon •Interoperable •Service Based •Adapve •Context Aware •Autonomous •Distributed

device. To explore services provided by a device semantic match making mechanism is used (Fabisch and Henninger 2019). The various challenges associated with discovery includes interoperability (Nitti et al. 2017), security (Zhang et al. 2017), and unique addressing (Cheng et al. 2015). • Resource Management: To provide QoS, it is required to deploy certain module on IoT platform to manage resources. Resources need to be monitored, scheduled and allocated in a fair, conflict free manner (Razzaque et al. 2015) by the platform. To carry out this task platforms must maintain data about device battery time, memory usage, processing power, and other relevant information for efficient resource management (da Cruz et al. 2018). • Data Management: Data management plays a vital role in any application development. In case of IoT, data refers to the data sensed by the IoT devices. IoT platform is required to provide data management services which include data acquisition, data processing, querying and visualisation. • Data Acquisition: IoT platforms acquire data sensed by devices from various sources. The sensed data can be structured or semi-structured (Cheng et  al. 2015). Following are the main sources of data captured by the platform (Santana et al. 2018; Khan et al. 2015): –– Real time data about physical devices such as traffic, city maps, citizen’s data. –– Data available in form of software such as libraries, codes, documents. –– Historically stored data in the form of logs, historic actions.

48









• •

P. Agarwal and M. Alam

The challenges associated with designing data acquisition are addressing, scalability, configuration, security and interoperability of objects (Apolinarski et al. 2014). Data Cleaning: This data acquired through sensors may be incomplete, inconsistent, noisy or irrelevant. To increase the reliability of the system it is required to be cleaned and preprocessed for further execution. The algorithms for anomaly detection (Cheng et al. 2015), maintaining semantic consistency, data normalization (Silva et al. 2018), and data filtering (Petrolo et al. 2014; Filipponi et  al. 2010) need to be designed. Some platforms even employ strategies that allow users to collect data of interest only by filtering irrelevant data through algorithms (Soldatos et al. 2015). Data Storing: It deals with storing the data acquired through the sensors. IoT platform is mainly responsible for managing huge volume and variety of IoT sensed data. Mainly relational databases and No SQL databases are used for storage (Fahmideh and Zowghi 2020). Relational databases are used for structural, transactional data. No SQL databases are used for dynamically changing schema. Some of the popular choices of No SQL databases are Hadoop, CouchDB, HBases, and MangoDB. Data Processing: This includes the power of IoT platform to analyses sensor data for meaningful inferences. Data is usually analyzed in two modes: Real time analytics or Streaming mode, and batch processing or historical analysis (Al-­ Fuqaha et al. 2015). In real time analytics the concern is over timely efficient processing of fast running IoT stream data. Some of the technologies that support fast streaming data are Apache Spark, Storm. Historic analysis deals with processing batch of historically stored sensor data and finding inferences in it. Data mining techniques such as classification, regression and clustering can be done on data to find classes, groups or abnormalities, trends in acquired sensor data. Query Processing: It is the process for querying the data stored in the IoT platform. Querying mechanism also uses publish/subscribe mechanism like data acquisition (Cheng et  al. 2015). The query can be a simple query or complex query aggregating data from multiple databases. Data Visualization: in response to users query on stored data, visualization techniques provide graphical view of the analysis results. Results can be in form of dashboards, maps or reports. Event Processing: The function of the platform is to process event successfully. Event is nothing, but a change in environment, which is captured by sensor, and requires certain action to be taken over it. Usually event processing is required to be carried out in real time (Cretu 2012). The platform should provide flexibility to the users in terms of executing own code and define event conditions (Sarhan 2019). The main challenges associated with event processing in IoT platform is to combine data from multiple sensor stream in different forms and combine them for one common goal. Events are usually represented in form of ontologies enabled by metadata for their correct interpretations (Fahmideh and Zowghi 2020).

3  Open Service Platforms for IoT

49

• Code Management: IoT platform plays an important role in deploying code for IoT applications. For this code allocation, and migration services are required (da Cruz et  al. 2018). Code allocation deals with selecting appropriate sensor devices and executing code on that particular sensor device. Code migration deals with providing portable facilities for migrating code on different programming services. Non –functional requirements: The various non-functional requirements are described as follows: • Scalability: The IoT platform needs to accommodate large number of devices. The IoT platform must provide features that can add any number of devices, and can remove any number of devices (Al-Fuqaha et al. 2015). Adding large number of devices must maintain QoS (da Cruz et al. 2018) • Timeliness: Event processing mainly rely on timely execution on data. Most of the events require real time processing. Real time means execution of data quickly, without any delay. The time required for processing is critical and is determined by different applications processing them. • Availability: For critical applications 24X7 availability of the platform is must. The platform must be available for services, even it is facing internal failures (Razzaque et al. 2015). The recovering time from failure should be very small, not affecting the system performance. Reliability and availability both are important factors and deals with fault tolerance. • Security and privacy: Another important requirement of IoT platform is to provide security to the user’s data. Most of the applications require to store user’s personal data and information such as GPS cation, password, etc. Proper security mechanisms need to be deployed to protect user’s information during transmission and storage to protect it from malicious attacks. Proper security and privacy preserving mechanisms need to be deployed at both functional and non-­functional level (Al-Fuqaha et al. 2015). • Ease of deployment: IoT platforms are used for end user application deployment. Most of the time they are used by application developers to integrate their own device. So, it is required from platform to be user friendly, to be easily integral with device, and must not require lot of expertise. Must be easy to install and set up.

3.3.2  Architectural Requirements Architectural requirements supports application development. It deals with the requirements which can ease the application development task, such as programming abstraction from hardware implementation, inbuilt API’s, libraries. The major architectural requirements are as follows:

50

P. Agarwal and M. Alam

• Programming Abstraction: The application developing programming interface of an IoT platform must be able to hide the internal working of the system. At the time of designing the IoT platform, it is required to design the level of abstraction to be provided to the different level of users. • Interoperable: An IoT platform must be able to interoperate heterogeneous devices, technologies and diverse applications. These heterogeneous devices must be able to communicate with each other, exchange information, and can work collaboratively to achieve final goal. Interoperability can be achieved at network, semantic, and syntactic level (Fahmideh and Zowghi 2020). • Service based: IoT platforms must support service-based framework, where new service interfaces for diverse applications can be easily added, without affecting the underlying hardware interfaces. Service oriented framework provides flexible architecture for building diverse applications. • Adaptive: IoT platform architecture should be capable of adapting itself to changing environment. IoT applications usually work in dynamic environment. So, it is required that platform should incorporate this dynamicity in its architecture. • Context Aware: Context awareness adds value to the information sensed by the IoT devices. Context awareness means IoT device must be capable of capturing user’s information and device information for providing more meaningful inferences. • Autonomous: All devices in IoT environment must work in self-governing mode. They must be able to work autonomously in collaboration with other devices in the network without human intervention (Gubbi et  al. 2013; Wang et al. 2010). Automaticity can be provided by embedded intelligence, analytics, and autonomous agents (Guo et al. 2011). • Distributed: In order to support distributive applications like transportation, traffic. The IoT platform must be able to provide distributed, decentralised processing. Platform must support functions that can be performed in physically distributed infrastructure environment.

3.4  Case Study of IoT Service Platforms The market is overwhelmed with number of IoT middleware platforms. Out of hundreds of IoT platforms, this chapter presents case study of four most popular platforms namely: Amazon Web Service (AWS) IoT, Microsoft Azure IoT, Google Cloud Platform, IBM Watson IoT.  Each one of the following is discussed with requirements viewpoint.

3  Open Service Platforms for IoT

51

3.4.1  Amazon Web Service (AWS IoT) AWS IoT platform is developed by Amazon Web Services (AWS IoT 2019). It can connect millions of IoT devices through secure gateway. It is an integration of large number of middleware technologies working collaboratively. AWS has lot of inbuilt technologies supporting integration with heterogeneous devices, capturing data, securely transmission on cloud, support for device authorization and authentication, analytics tools. Besides these, AWS also supports number of third party applications. The technologies supported by AWS, to satisfy service and architectural requirements of the platform are given below. Functional Requirements  AWS follows publish subscribe mechanism for resource discovery. Each device connects to AWS via. gateway through secure channel protocols. Data is then passed through Rules Engine, which transforms data and pass it to the cloud services. At cloud end, Dynamo DB for NoSQL storage is deployed, lambda functions are used for event management. AWS supports dashboards for visualisation of data. Non-functional Requirements  AWS can scale millions of devices. It interconnects devices irrespective of their underlying architecture using API’s and SDK’s. It supports large number of programming language, and supports code migration. It can integrate with large number of Operating system through command line interface. Each device is authenticated and authorised using certificate, and is assigned unique ID. Data during transmission is secured through encryption and decryption. Overall AWS is easy to deploy, and uses pay as you use policy. Architectural Requirements  AWS provides programming abstraction, as user can program in any language. AWS has support for large number of interfaces. It has high fault tolerance and availability. Things Shadow maintains data related to device such as identification, location. It has centralized architecture, but supports distributed processing with Elastic MapReduce. The different features corresponding to requirements are summarized in Table 3.1.

3.4.2  Microsoft Azure IoT Microsoft Azure IoT is developed by Microsoft (Azure IoT | Microsoft Azure 2019). It is considered to be only hybrid cloud service solution. Unlike, AWS IoT it can support pre-configured solutions. Azure has one of the powerful artificial intelligence support engine. It has number tools and technologies for securing capturing data, storing, and processing data. In response to various service and architectural requirements, various tools and technologies deployed are discussed below.

52

P. Agarwal and M. Alam

Table 3.1  AWS IoT features Service requirements Functional requirements Resource Publish and subscribe mechanism via. Message broker on device gateway discovery through MQTT or HTTP protocol Resource Things registry is used to manage resources allocated with each device. management Elastic load balancing to manage load in network Data management Amazon simple storage service (S3) provides scalable storage, dynamo DB provides NoSql databases, lambda for virtualization Query processing Rules engine using SQL language for message processing Data visualization Dashboards for visualization Event processing Amazon Kenesis for real time event processing. Simple notification service is used for event notifications. Lambda functions can trigger different events Code management Through in built SDK’s can code in any language Non-functional requirements Scalability Can scale billions of heterogeneous devices Timeliness Amazon Kenesis with Kenesis analytics is used for real time processing Availability 24 × 7 Security and Certificates for authentication, supports encryption decryption of data, user privacy can define its own security rules and policies Ease of Easy. With just registration deployment Architectural requirements Programming Command line Interface is compatible with number of OS such as abstraction windows, Linux, and OSX. AWS SDK’s allow user to program in any language Interoperable Rules engine makes it interoperable with other services Service based Support for large number of API’s and SDK’s make architecture service based Adaptive Not very adaptive friendly platform ContextThings shadow maintain current state information of device Awareneness Autonomous Each device can autonomously collaborate with other devices to exchange data using its unique identification number Distributed It has centralized architecture, but uses elastic MapReduce for processing

Functional Requirements  Each device registers itself with azure directory service, which provides it a unique identification. Load balancing by azure platform is carried out both at local as well as global level. Supports both NoSQL storage for semi structured data, as well as support for SQL storage. It supports large azure analytics engine. Number of powerful tools for visualisation are also available. Supports real time event processing and code management. Non-functional Requirements  Azure can be scaled to large number of devices. Supports both real time as well as historic processing of data. In order to improve

3  Open Service Platforms for IoT

53

availability, back up mechanism is deployed to make system fault tolerant. To provide device authentication, two way secure protocol is deployed. Architectural Requirements  Azure can be deployed either in windows, or linux platform. It has support for number of programming languages such as python, java, PHP, Node.js, etc. To interconnect large number of heterogeneous devices with different underlying architecture, number of inbuilt SDK’s and APIs are available. Mean Time to Failure is quite low. Supports decentralized serverless architecture. The service and architectural requirements of Microsoft azure are summarized in Table 3.2.

3.4.3  Google Cloud Platform Google Cloud Platform (GCP) is developed by Google. It can connect heterogeneous IoT devices. It can support lightweight applications. It is considered to be server less architecture. GCP has support for large number of analytics libraries like tensor flow, etc. GCP is one of the powerful emerging IoT platform supporting large number of features, including support mobile applications. Various features supporting functional, non-functional and architectural requirements are given below: Functional Requirements  GCP uses publish subscribe mechanism for resource discovery. A registry of device is maintained. It supports BigTable storage solution, and firebase solution for real time processing of the event. Supports large number of programming languages interface. Non-functional Requirements  Can scale millions of devices. In order to make system fault tolerant, it deploys back up mechanism. Security and privacy is provided by authorisation and authentication of device. Firebase solution deals with the timeliness of the event execution. It supports large number of machine libraries like tensorflow to provide intelligent solutions. Architectural Requirements  GCP provides good programming abstraction by supporting large number of programming languages. Context aware applications are easy to deploy as each device in the network can be tracked by unique ID, and location. Suitable for lightweight mobile apps as well as client web applications. It uses centralized architectural approach. The Service and architectural requirements of Google Cloud Platform are summarized in Table 3.3.

54

P. Agarwal and M. Alam

Table 3.2  Microsoft Azure features Service requirements Functional requirements Resource Not much support for resource discovery discovery Resource Azure active directory. Supports both global level and local level load management balancing Data management Support storage using Cosmos DB and processing, support for in-motion analytics. Data can also be stored in blob storage and Postgre storage Query processing Azure analytics through SQL syntax Data visualization Dashboards, power BI, web apps Event processing Support for predictive analytics, user can set a thresholds and alert limits through event hub Code management Supports code migration Non-functional requirements Scalability Scalability of ten million devices per instance Timeliness Support both historic and real time processing. Supports stream analytics Availability 24 × 7. supports fault localization. Maintains backup to improve reliability Security and Device authentication through two way secure protocol privacy Medium level difficulty Ease of deployment Architectural requirements Can scale with number of programming languages and command line Programming abstraction interface through inbuilt tools and technologies Interoperable Agent libraries, SDK’s allow interoperability within heterogeneous systems Service based Provides certain pre- configured solutions. Support number of architectures like event driven, microservices, n-tier, web-queue, and worker Adaptive MTTR is very low ContextAzure active directory awareness Autonomous Heterogeneous devices can work autonomously in collaboration to each other Distributed Decentralized serverless architecture

3.4.4  IBM Watson IoT IBM Bluemix platform is recently named as IBM Watson IoT platform. It supports large number of cognitive analytics, and considered one of the popular platform for researchers (IBM Knowledge Center 2019). The various technologies provided by IBM Watson in view of functional, non-functional and architectural requirements are as follows: Functional Requirements  Each device in the network is connected via. Message broker through secure gateway. Each device is assigned a unique organisation ID. It supports both SQL databases and NoSql databases. Cognitive engine is deployed

3  Open Service Platforms for IoT

55

Table 3.3  Google Cloud platform features Service requirements Functional requirements Resource discovery Uses pub/ sub scheme. Devices register through device id. Maintains registry of devices Resource Devices managed through device drivers and protocol bridge. Deploy management resource allocation Data management Cloud BigTable is used for data storage. Firebase database for real time processing. Supports both historic and real time data processing Query processing BigQuery is use for query processing Data visualization Dashboards Event processing Firebase database for real time event time processing Code management Support number of programming languages and libraries Non-functional requirements Scalability Can scale millions of devices Timeliness Real time processing though firebase Availability 24 × 7. Fault tolerance through backup Security and Authentication, authorization of device privacy Easy Ease of deployment Architectural requirements Number of programming languages Programming abstraction Interoperable Number of libraries, lightweight libraries, APIs, SDK Service based Supports mobile apps, and end to end applications Adaptive Not very adaptive ContextWith each device there is unique id, state information, telemetry Awareneness Autonomous Each device with unique id can send stream of telemetry Distributed Centralized approach

for query processing. Dashboards and Jupyter notebook for visualization. Supports both historic and real time event processing. Non-Functional Requirements  can scale millions of devices. Kafka REST APIs are used to provide timeliness of event processing. Provide authentication, authorisation mechanism through protocols, with support for risk management. Platform is quite easy to deploy, making it popular choice among researchers. Architectural Requirements  Supports large number of programming language. Usually Node Red editor is used for programming. Suitable for deploying context aware apps. Supports large number of SDKs. Can be used to build lightweight as well as web apps. It has centralized architecture with support for distributed ­processing. The service and architectural requirements for IBM Watson are summarized in Table 3.4.

56

P. Agarwal and M. Alam

Table 3.4  IBM Watson features Service requirements Functional requirements Resource Message broker for secure device registration via. Gateway discovery Resource Each device registers through unique organization id management Data management Cloudant NoSql DB for real time processing. Data lake for SQL storage. DB2 for long term schema storage Query processing Has cognitive engine for providing data analytics capability. Supports machine learning. Supports both predictive and prescriptive analytics Data visualization Dashboards, support jupyter notebook Event processing Support historic and real time event processing Code management Supports code migration Non-functional requirements Scalability Can scale millions of devices Timeliness Real time processing through no SQL event streaming and Kafka REST APIs Availability High availability Security and Unique authentication code for device identification. Supports risk privacy management. Secure communication via. Protocols Ease of Easier to deploy deployment Architectural requirements Programming User can choose own programming language and architecture. Red node abstraction visual programming editor Interoperable Large number of SDK’s Service based Framework supports client side application, mobile micro apps Adaptive Can adapt to changing environment ContextStores device id, last activity, geographic location awareneness Autonomous Each device has unique social id and can autonomously collaborate with each other Distributed Centralized architecture with support for distributed processing

3.5  Challenges and Open Research Problems The IoT Platforms, still have lot of research challenges that need to be addressed in future research. The various challenges and open research problems related to IoT platforms are discussed below: • Need for improved and more accurate models for resource discovery. IoT platform needs to address large number of request in timely and accurate manner. The present registry methods such as distributed, hybrid, or probabilistic does not fully cater this requirement (Teixeira et  al. 2011). So, there is a need for designing better resource discovery models.

3  Open Service Platforms for IoT

57

• Need for more efficient resource scheduling and management policies. Conflict of resources among IoT devices is a very common scenario. At present very few IoT platforms deploy strategies for resource conflict resolution. So, this is one of the area which need to be addressed in the future. • More focus on support for data filtering. Data aggregation and filtering is one of the major steps in analysis. Most of the platforms provide data aggregation solution, but very few support on filtering of relevant and irrelevant data. • Support for data compression. IoT generates huge amount of data, requiring large amount of storage. Proper data compression solutions can be deployed to reduce the amount of storage required. • Support for changing business logic or IoT environment. Business logic and requirements keep on changing with time. Proper code allocation and migration strategies need to be deployed with support for firmware update. • Support for Hard real time processing is required. Soft real time processing of events is done efficiently IoT.  But, still addressing Hard time event is a challenge. • Seamless replacement of modules to provide reliability in case of faults and failures is still a challenge. • Service provising in case of failure to improve availability of a platform in seamless manner needs to be addressed in future research. • Human intervention is required for deployment of devices on IoT platform. There is a need for pre-configured solutions to address this issue and reduce human efforts, making things automated and ease for deployment. • Need to address syntactic and semantic interoperability. Most of the IoT platforms provide hardware interoperability with lesser focus on syntactic and semantic interoperability. • Need for more dynamic rules and policies for making deployed system adaptable to run time environment.

3.6  Conclusion In IoT scenario heterogeneous devices are connected and work collaboratively to achieve a particular goal. At the lowest layer, lies the resource constrained sensors, that senses the environment. To convert this information into meaningful insights, the data through sensors need to be stored and processed. All these packages need to be consolidated into one platform called as IoT platform. Number of vendors are available in the market with different requirements and specifications. Choosing the right one according to the need of application is the major key factor in successful deployment of application. The IoT platforms have come a long way, still they need to work on certain aspects to provide better functionality like security, resource provision, making more user friendly, easy device discovery, deployment and integration, better storage and resource management policies.

58

P. Agarwal and M. Alam

References Agarwal, P., & Alam, M. (2019). Investigating IoT middleware platforms for smart application development. arXiv preprint arXiv, 1810, 12292. Alam, M. (2012, December). Cloud Algebra for handling unstructured data in cloud database management system. International Journal on Cloud Computing: Services and Architecture (IJCCSA), 2(6) ISSN: 2231  – 5853 [Online], 2231  – 6663 [Print], https://doi.org/10.5121/ ijccsa.2012.2603, Taiwan. Alam, M., & Shakil, K. A. Presented “Big Data analytics in Cloud environment using Hadoop. In International conferences on Mathematics, Physics & Allied sciences-2016, March 03–05, 2016, Goa. Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., & Ayyash, M. (2015). Internet of things: A survey on enabling technologies, protocols, and applications. IEEE Communications Surveys & Tutorials, 17(4), 2347–2376. Ali, S.  A., Affan, M., & Alam, M. (2019). A study of efficient energy management techniques for Cloud Computing environment. 2019 9th International conference on Cloud Computing, Data Science & Engineering (Confluence) (pp. 13–18), Noida, India. https://doi.org/10.1109/ CONFLUENCE.2019.8776977. Apolinarski, W., Iqbal, U., & Parreira, J. X. (2014, March). The GAMBAS middleware and SDK for smart city applications. In 2014 IEEE International conference on pervasive computing and communication workshops (PERCOM WORKSHOPS) (pp. 117–122). IEEE. Asghari, P., Rahmani, A. M., & Javadi, H. H. S. (2019). Internet of things applications: A systematic review. Computer Networks, 148, 241–261. Ashton, K. (2009). That ‘internet of things’ thing. RFID Journal, 22(7), 97–114. Atzori, L., Iera, A., & Morabito, G. (2010). The internet of things: A survey. Computer Networks, 54(15), 2787–2805. AWS IoT. (2019). Retrieved September 12, 2019, from https://docs.aws.amazon.com/iot/index. html Azure IoT | Microsoft Azure. (2019). Retrieved September 12, 2019, from https://azure.microsoft. com/en-in/overview/iot/ Cheng, B., Longo, S., Cirillo, F., Bauer, M., & Kovacs, E. (2015, June). Building a big data platform for smart cities: Experience and lessons from santander. In 2015 IEEE International congress on Big Data (pp. 592–599). IEEE. Cretu, L. G. (2012). Smart cities design using event-driven paradigm and semantic web. Informatica Economica, 16(4), 57. da Cruz, M. A., Rodrigues, J. J. P., Al-Muhtadi, J., Korotaev, V. V., & de Albuquerque, V. H. C. (2018). A reference model for internet of things middleware. IEEE Internet of Things Journal, 5(2), 871–883. Fabisch, M., & Henninger, S. (2019). ESPRESSO–systemic standardisation approach to empower smart cities and communities. Smart Cities in Smart Regions, 2018, 115. Fahmideh, M., & Zowghi, D. (2020). An exploration of IoT platform development. Information Systems, 87, 101409. Filipponi, L., Vitaletti, A., Landi, G., Memeo, V., Laura, G., & Pucci, P. (2010, July). Smart city: An event driven architecture for monitoring public spaces with heterogeneous sensors. In 2010 Fourth International Conference on Sensor Technologies and Applications (pp.  281–286). IEEE. Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013). Internet of Things (IoT): A vision, architectural elements, and future directions. Future Generation Computer Systems, 29(7), 1645–1660. Guo, B., Zhang, D., & Wang, Z. (2011, October). Living with internet of things: The emergence of embedded intelligence. In 2011 International conference on internet of things and 4th international conference on cyber, physical and social computing (pp. 297–304). IEEE.

3  Open Service Platforms for IoT

59

IBM Knowledge Center. (2019). Retrieved September 12, 2019, from https://www.ibm.com/ suport/knowledgcenter/SSQP8H/iot/overview/architecture.html Khan, Z., Anjum, A., Soomro, K., & Tahir, M. A. (2015). Towards cloud based big data analytics for smart future cities. Journal of Cloud Computing, 4(1), 2. Khan, S., Shakil, K. A, & Alam, M. (2017). Cloud based Big Data Analytics: A survey of current research and future directions, Big Data Analytics (pp 629–640). Springer, Print ISBN: 978-­ 981-­10-6619-1, Electronic ISBN: 978–981-10- 6620-7. Nitti, M., Pilloni, V., Giusto, D., & Popescu, V. (2017). Iot architecture for a sustainable tourism application in a smart city environment. Mobile Information Systems, 2017. Overview of Internet of Things | Solutions | Google Cloud. (2019). Retrieved September 12, 2019, from https://cloud.google.com/solutions/iot-overview Petrolo, R., Loscri, V., & Mitton, N. (2014, August). Towards a smart city based on cloud of things. In Proceedings of the 2014 ACM international workshop on wireless and mobile technologies for smart cities (pp. 61–66). ACM. Razzaque, M. A., Milojevic-Jevric, M., Palade, A., & Clarke, S. (2015). Middleware for internet of things: A survey. IEEE Internet of Things Journal, 3(1), 70–95. Santana, E. F. Z., Chaves, A. P., Gerosa, M. A., Kon, F., & Milojicic, D. S. (2018). Software platforms for smart cities: Concepts, requirements, challenges, and a unified reference architecture. ACM Computing Surveys (CSUR), 50(6), 78. Sarhan, A. (2019). Cloud-based IoT platform: Challenges and applied solutions. In Harnessing the Internet of Everything (IoE) for accelerated innovation opportunities (pp.  116–147). Hershey: IGI Global. Silva, B. N., Khan, M., & Han, K. (2018). Internet of things: A comprehensive review of enabling technologies, architecture, and challenges. IETE Technical Review, 35(2), 205–220. Soldatos, J., Kefalakis, N., Hauswirth, M., Serrano, M., Calbimonte, J. P., Riahi, M., et al. (2015). Openiot: Open source internet-of-things in the cloud. In Interoperability and open-source solutions for the internet of things (pp. 13–25). Cham: Springer. Teixeira, T., Hachem, S., Issarny, V., & Georgantas, N. (2011, October). Service oriented middleware for the internet of things: A perspective. In European conference on a service-based internet (pp. 220–229). Berlin/Heidelberg: Springer. Tiwana, A. (2013). Platform ecosystems: Aligning architecture, governance, and strategy. Oxford: Newnes. Wang, H., Zhou, X., Zhou, X., Liu, W., Li, W., & Bouguettaya, A. (2010, December). Adaptive service composition based on reinforcement learning. In International conference on service-­ oriented computing (pp. 92–107). Berlin/Heidelberg: Springer. Yaqoob, I., Ahmed, E., Hashem, I. A. T., Ahmed, A. I. A., Gani, A., Imran, M., & Guizani, M. (2017). Internet of things architecture: Recent advances, taxonomy, requirements, and open challenges. IEEE Wireless Communications, 24(3), 10–16. Zhang, K., Ni, J., Yang, K., Liang, X., Ren, J., & Shen, X. S. (2017). Security and privacy in smart city applications: Challenges and solutions. IEEE Communications Magazine, 55(1), 122–129.

Part II

Solutions and Enablers for IoT

Chapter 4

Resource Management Techniques for Cloud-Based IoT Environment Syed Arshad Ali, Manzoor Ansari, and Mansaf Alam

Abstract  Internet of Things (IoT) is an Internet-based environment of connected devices and applications. IoT creates an environment where physical devices and sensors are flawlessly combined into information nodes to deliver innovative and smart services for human-being to make their life easier and more efficient. The main objective of the IoT devices-network is to generate data, which are converted into useful information by the data analysis process, it also provides useful resources to the end-users. IoT resource management is a key challenge to ensure the quality of the end user’s experience. Many IoT smart devices and technologies like sensors, actuators, RFID, UMTS, 3G, and GSM etc. are used to develop IoT networks. Cloud Computing plays an important role in these networks deployment by providing physical resources as virtualized resources consist of memory, computation power, network bandwidth, virtualized system and device drivers in secure and pay as per use basis. One of the major concerns of Cloud-based IoT environment is resource management, which ensures efficient resource utilization, load balancing, reduces SLA violation, and improve the system performance by reducing operational cost and energy consumption. Many researchers have been proposed IoT based resource management techniques. The focus of this paper is to investigate these proposed resource allocation techniques and finds which parameters must be considered for improvement in resource allocation for IoT networks. Further, this paper also uncovered challenges and issues of Cloud-based resource allocation for IoT environment. Keywords  IoT · Cloud computing · Resource allocation · Parameters · Fog computing

S. A. Ali (*) · M. Ansari · M. Alam Department of Computer Science, Jamia Millia Islamia, New Delhi, India e-mail: [email protected]; [email protected]; [email protected] © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_4

63

64

S. A. Ali et al.

4.1  Introduction The Internet of Things (IoT) is a set of connected smart device and sensors over the Internet. These devices are connected using the wired/wireless network technologies to communicate and transfer data from one node to another (Al-Fuqaha et al. 2015; Horrow and Sardana 2012; Pourghebleh and Navimipour 2017; Yang et al. 2013). The things in the IoT infrastructure network are sensors, smart devices, sensor data, software agents and human beings (Yan et  al. 2014). These networked independent devices make local network and connected to the global network to share information with others in real-time to realize that the Things are connected into the real world and connect all devices (Alaba et al. 2017; Lee and Lee 2015; Mattern and Floerkemeier 2010). In the cyber-physical ecosystem, each edge-node is supposed to an IoT device which can dynamically cooperate with other devices in the network to execute one or more user’s tasks allocated to the system network. Resources like processing power, storage, network bandwidth, RAM are usually limited in these IoT networks, though these infrastructures and computing resources are provided by the Cloud service providers. IoT devices produce a huge amount of real-time data from the sensors. Cloud storage is used for storing these real-time data on different local networked data centres, which upload these data to the global networked data centres to allow access for all globally situated smart devices (Bassi et al. 2013). In this paper, the author studied various resource allocation techniques for Cloud-based IoT system. Classification of these techniques has been done based on parameters like QoS, context, cost, energy consumption and SLA. Furthermore, the author also discussed various parameters of resource allocation techniques of the IoT system.

4.2  Basic Concepts of IoT, Cloud and Resource Management 4.2.1  What Are the Basic Elements of the IoT Environment? Internet of things offers numerous advantages and services to the users. Therefore, to use them correctly, some elements are needed. The IoT elements will be discussed in this section. Figure  4.1 shows the elements required to provide IoT functionalities. 4.2.1.1  Identifiers Within the network, it offers an explicit identity for each object. In identification, two processes exist naming and addressing. The naming relates to the object’s title whereas the addressing explains the address of an object. These two processes are very different even though two or more objects could have the same name, but they

4  Resource Management Techniques for Cloud-Based IoT Environment

65

Fig. 4.1  Basic elements of the IoT environment

are always different and unique. There are several approaches accessible which expedite the naming of objects in the network, such as ubiquitous codes (uCode) and electronic product codes (EPCs) (Koshizuka and Sakamura 2010). Using IPv6, each object has a unique address. First, IPv4 was used to allocate the address, but due to a huge amount of IoT devices, it was unable to address the need. IPv6 is therefore used because it uses a 128-bit addressing system. 4.2.1.2  Sensing Devices This involves acquiring information from the environment and transferring it to a local, remote or Cloud-based database as an instance of the IoT sensor. We can identify intelligent devices, portable sensors or actuators. The collected information is transmitted to the storage medium. Numerous detection devices to collect data on objects such as RFID tags, actuators, portable sensors, smart sensors, etc. 4.2.1.3  Communications Devices To achieve smart services, IoT communication techniques communicate heterogeneous artifacts. One of the main goals of the Internet of things is Communication in which various devices connect and communicate with each other. In the communication layer, devices can transfer and deliver messages, documents and other information. There are many methods which facilitate communication, for example

66

S. A. Ali et al.

Bluetooth (Mcdermott-Wells 2004), radio frequency identification (RFID) (Want 2006), long term evolution (LTE) (Crosby and Vafa 2013), Wi-Fi (Ferro and Potorti 2005) and near-field communication (NFC) (Want 2011). 4.2.1.4  Compute Devices The computation of the data collected by the objects is achieved using sensors. It is used to develop processing in IoT applications. Raspberry Pi, Arduino, and Gadgeteer are utilized for hardware platforms, whereas the operating system plays a significant role in the processing of software platforms. There is various kind of operating systems are used, including Lite OS (Cao et al. 2008), Riot OS (Baccelli et al. 2013), Android, Tiny OS (Levis et al. 2005) etc. 4.2.1.5  Services IoT Applications provide four types of services (Gigli and Koo 2011; Xing et al. 2010). The first service that is associated with an identity. It is used to acquire the identity of the objects which sent the request. The aggregate of information is another service aimed at collecting all the data about the objects. The aggregation service also performs the processing. The third service refers to co-operative service that makes decisions based on the information gathered and transfer suitable rejoinders to the devices. The last service is the pervasive service, which is used to replace devices immediately without rigidity in terms of time and place. 4.2.1.6  Semantics They are the IoT’s concern to facilitate the consumers who perform their tasks. To fulfil its responsibilities, it is the most important component of IoT. It performs as the IoT’s brain. It accepts all the information and makes the appropriate decisions to send responses to the devices. The key techniques used in each IoT component are presented in Table 4.1. Table 4.1  Key technologies used by IoT components IoT components Identification Sensing Communications Computation Services Semantics

Main technologies IPv4, IPv6 Electronic product code(eCodes), ubiquitous code (uCode), Actuators, Sensors, Wearable Sensing Devices, RFID Tags. Wireless Sensor Network (WSN), Long Term Evolution (LTE), Bluetooth, Near Field Communication (NFC), Radio Frequency Identification (RFID), Intel Galil Operating System, Arduino, Raspberry Pi. Collaborative-Aware, Ubiquitous Identity-Related, Information Aggregation EXI, OWL, RDF

4  Resource Management Techniques for Cloud-Based IoT Environment

67

4.2.2  What Are the Various IoT Architecture Frameworks? An IoT architecture can be viewed from three viewpoints: Things-oriented, Internet-­ oriented and semantic-oriented (Atzori et al. 2010). In things-oriented viewpoint, the intelligent autonomous smart devices are connected to each other using NFC and RFID technologies for specific daily life applications. Internet-oriented viewpoint focuses on how these smart devices connect to the internet using unique identification (IP addresses) and standard communication protocols to facilitate the global connections among these applications based smart devices. In the semantic viewpoint of IoT architecture, the data generated by the IoT devices are used to generate useful information and handle the architectural modelling problems efficiently using these produced information (Marques et al. 2017). In the opinion of most researchers on conventional IoT architecture, it is considered at three layers: -Perception Layer, Network Layer, Application Layer. In addition, some researchers have analyzed another layer also included in the latest IoT architecture, which is a support layer located between the network layer and the application layer. Multi-tier architecture of IoT is displayed in Fig. 4.2. The support layer consists of fog computing and Cloud Computing. In this section, we define seven-tier architecture: collaboration and processes, applications, data abstraction, data accumulation, edge computing, connectivity, physical devices and controller. The basic layered architecture of IoT and its components in each layer have been depicted in Fig. 4.3.

Fig. 4.2  Multi-tier IoT architecture

68

S. A. Ali et al.

Fig. 4.3  Basic architecture of IoT

4.2.2.1  Layer 1: Physical Devices and Controllers It is a layer of perception or hardware that collects and sends information from the physical world to the next layer. This layer includes objects and physical sensors. Basically, this layer is intended to detect various objects and collect environmental data such as humidity, temperature, water quality, pressure, air quality, motion detection etc. Controllers and Physical devices can control multiple devices. These are the “things” in the IoT containing a wide range of endpoint devices for sending and receiving information. The list of devices is already extensive today. It will be effectively unlimited as more devices will be added to the IoT over time. 4.2.2.2  Layer 2: Connectivity This layer is used to interact with various IoT components via interconnecting systems such as switches, gateways as well as routers. In addition, it transfers data collected strongly from the sensors to the top layer for processing. It includes transferring data from physical devices to Cloud or other devices, which may be in the form of ZWave, SigFox Zigbee or Bluetooth. This layer extends to Cloud transport services from the “intermediate” of an Edge Node device.

4  Resource Management Techniques for Cloud-Based IoT Environment

69

4.2.2.3  Layer 3: Edge Connecting The next phase is Edge Computing, or more appropriately “Cloud Edge” or “Cloud Gateway”. Layer 3 requires data from the connectivity phase and makes it suitable for archiving and higher-level processing information. The processing elements in this layer work with a large volume of data that could transform some facts to moderate data size. 4.2.2.4  Layer 4: Data Accumulation They must be stored after the data is accumulated. It is important where information is stored. While some data may be stored to the limit, most data will have to be delivered to the Cloud. Big Data machines will be able to exploit their computing power and analyze the data. The main objective of this layer is to store the data of Phase 3. Acquire and store a large amount of data and place them in the warehouses so that they are accessible from the upper layers. As a result, it simply modifies event-based data in query-based processing information for the higher layer. This layer can be deployed in SQL or requires a more sophisticated Hadoop and Hadoop file system, Mongo, Cassandra, Spark or other NoSQL solutions. 4.2.2.5  Layer 5: Data Abstraction This layer combines data from different sources and converts the stored data to the appropriate application format in a manageable and efficient way (Atlam et  al. 2017). A main component of the large-scale high-performance implementation architecture is a publishing/subscription software framework or a data distribution service (DDS) to simplify data movement between edge computing, data accumulation, application layers and processes. Whether it’s a high-performance service or a simple message bus, this infrastructure simplifies deployment and improves performance for all applications, except for the simplest. 4.2.2.6  Layer 6: The Application Layer It applies to information elucidation from different IoT applications. It comprises several IoT applications, like medical care, smart city, smart network, smart agriculture, smart building, connected car etc. (Stallings 2015). This phase is self-­ explanatory where the application logic of the control plan and the data plan are performed. Process optimization, logistics, statistical analysis, control logic, Monitoring, alarm management, consumption models are just certain cases of IoT applications.

70

S. A. Ali et al.

4.2.2.7  Layer 7: Collaboration and Processes At this phase, application processing and collaboration are presented to users, and the processed data in the lower layers are incorporated into commercial purposes. It encompasses collaboration, people, businesses, and decision-making processes based on IoT-derived information. This layer classifies people who can collaborate and communicate to use IoT data proficiently. It delivers additional features, such as the creation of commercial graphics and models and other data-based recoveries from the application layer. It also helps executives make accurate business decisions based on data analysis (Muntjir et al. 2017).

4.2.3  How Cloud Computing Supports IoT Infrastructure? National Institute of Standards and Technology (NIST) defined Cloud Computing as a paradigm that enables global, appropriate and on-demand self-serviced shared pool of computing resources (network, storage, applications and services) to the end-user on pay per use basis with minimal effort or service provider communication (Mell and Grance 2011). Cloud Computing is used rapidly by IT industries and professional for the deployment of their projects due to minimal cost and rapid elastic characteristics of Cloud- services (Varghese and Buyya 2018). Cloud Computing is an Internet-based technology that facilitates both service-consumers and providers by its essential features of On-demand self-service, broad network access, resource pooling, rapid elasticity and measured services (Zhang et al. 2010). Cloud Computing has three service models and four deployment models. Cloud Service provider’s software running on Cloud set-up is used by the Cloud consumers in Software as a service (SaaS) model. Cloud consumers can deploy their applications on to the provider’s infrastructure using software design languages, libraries and tool provided by the Cloud infrastructure in Platform as a Service (PaaS) model. In Infrastructure as a Service (IaaS) model, a consumer can provide a pool of resources like storage, processing elements, network and another basic computing requirement for running any software and operating system provided by the Cloud provider as a virtual machine (VM) (Wang et  al. 2010). Cloud Computing also ­provides four deployment model as per the user priorities and features named as public, private, community and hybrid Cloud deployment models (Subashini and Kavitha 2011). Due to massive features and ability, Cloud Computing is used by many growing technologies, IoT is one among them. Integration of IoT with Cloud Computing has been intensively used by many real-life applications like smart cities, healthcare, agriculture industries, transportation, smart vehicles and many more (Malik and Om 2018). The current trends of technologies are moving towards the use of globally connected smart devices, which produces a huge amount of data that cannot be stored in locally due to limited storage capacity. The running fuel of all industries is the data, data is useless without its analytics to get useful information for industries future-plans and policies. The huge amount of data gathered by the

4  Resource Management Techniques for Cloud-Based IoT Environment

71

Fig. 4.4  How cloud support IoT applications

smart connected devices needs powered computing systems that should capable enough to compute this huge amount of data. The local systems are not capable enough to process these data for analytics. These limitations of huge storage space, computing power and network bandwidth can be overcome using pooled resources provide by Cloud Computing. Figure 4.4 illustrates how Cloud Computing supports IoT for smooth functionality. Cloud Computing infrastructure is used between the applications and IoT devices as a hidden layer, where all functionalities are hidden from the implementation of IoT based applications (Evans 2011).

4.2.4  Why Resource Allocation Is Important for IoT? Highly efficient maintained and cost-benefit network ensures the Quality of Service (QoS) standards. IoT architecture has various resources connected with the network. The resource allocation is an important aspect for QoS standards because of the efficient and effective allocation of resources in the IoT network. Resource ­allocation is also responsible for a high standard of security because in IoT architecture the data is divided into many data streams gathered from different sensors and different types of services are provided by the networked devices. The IoT networked resources consist of computing elements, storage and energy. Efficient Cloud resource allocation helps IoT networked devices to utilize these resources in an efficient and cost-effective way to improve system performance and productivity. IoT devices and resources are heterogeneous and globally distributed in nature, therefore, resource allocation and management are very important aspect of the IoT environment. The entity in the IoT network system can be an object, a human being or a place that is used to communicate between the IoT system application and the system user. These allocated objects are known as resources which can be

72

S. A. Ali et al.

categorized based on the information they communicate in the network. The threelayered IoT system mainly used are classified in IoT things, edge and Cloud infrastructure as described in Fig.  4.5. Allocation of resources by the system to accomplish the user tasks required various phases, first, the system selects the required resource from these three layers, then it selects the nodes which are capable to execute the user’s tasks and after that, it schedules the user’s task to the selected node for its execution. Finally, communication among these networked resources is done to accomplish the successful execution of the user’s task. Resource allocation has many aspects of resource discovery, resource provisioning, resource scheduling and resource monitoring as well. An effective Resource allocation supports standard Quality of Service (QoS), cost minimization, energy consumption reduction, increase resource utilization and moreover, it guaranteed the Service level agreement between the Cloud-based IoT system application providers and costumers where user’s requirements should be matched in an effective way. Resource allocation for IoT devices has many challenges due to heterogeneity and distance between the devices. A lot of research work has been done by the industries and academic researchers, but many challenges and issues related to IoT resource allocation are up till now untouched. Many researchers are working to solve these challenges and proposed new methods and algorithms for this. In this paper, the author systematically reviewed many proposed methods for resource allocation in Cloud-based IoT environment and classifies these techniques based on characteristics and resource allocation parameters improved by the technique.

Fig. 4.5  Three-tier architecture (IoT, Edge and Cloud)

4  Resource Management Techniques for Cloud-Based IoT Environment

73

Furthermore, the author discussed various parameters of IoT resource allocation and what parameters are still needed more improvements.

4.3  Related Works Many research papers have on been published related to survey and review of IoT technologies, wireless sensor network, integration of IoT and Cloud Computing and IoT architectures. But there is only one survey paper is present on the resource allocation in the IoT environment. Many researchers are working on various aspects of IoT architecture. The proposed methods and technologies that support the IoT networked system for smooth and cost-effective performance. No review paper exists for Cloudbased IoT resource allocation techniques. The author presented an important survey in (Lin et  al. 2017), that discussed various challenges and issues of IoT resource allocation from the architectural viewpoint. In this review paper, the author discussed the architectures and infrastructures that are responsible for resource allocation of the IoT environment. Resource scheduling and optimization techniques form the Cloud perspectives are not discussed in this paper. Another review paper is (Delicato et al. 2017) based on various IoT technologies and their challenges. In this paper, the author finds the relation between fog computing and IoT also discussed various existing resource allocation techniques in fog based IoT environment. The author discussed two main challenges in fog based IoT infrastructure, first is that the fog computing neither cares about the nearest node which provides computing resource nor data processing, it only cares about the minimum delay. The second issue the author found is that the resource allocation between the fog and IoT smart devices, because of the limited resource capacity of fog computing. To solve this problem of scarcity of fog resources, the author suggests the use of Cloud Computing above fog computing. In this paper, the author studied various Cloud-based IoT resource allocation techniques and classified these techniques in different groups. Further limitation and improvement were done in these techniques are also discussed and presented in the tabular form. Various resource allocation parameters have been listed out and defined. The author also discussed How much i­ mprovements have been done in these parameters and how much improvement still needed in remaining parameters.

4.4  C  lassification of Cloud-Based IoT Resource Management Techniques In this section, the author studied and classified Cloud-based IoT resource allocation techniques in various categories according to the feature provided by the technique under study. The detailed classification of IoT resource allocation techniques is described in Fig.  4.6. Various papers have been studied and categorized under these categories of IoT resource allocation.

74

S. A. Ali et al.

Fig. 4.6  Classification of IoT resource allocation techniques

4.4.1  SLA-Aware IoT Resource Allocation Service level agreement between the providers and consumer is very important in any service-oriented system. The SLA violation should be minimum to increase the profit of the service provider and satisfied the customer’s requirements. Many kinds of research have been done in this area of interest to reduce the SLA violation to improve the system acceptability among the costumers. Some research works are also done related to the SLA oriented resource allocation for IoT enable the system. In (Choi and Lim 2016), the author considered the penalty cost of SLA violation. A combinatorial auction method is proposed to reduce the penalty of SLA violation by calculating the provider’s profit and announce a winner to the user who gives maximum profit to the providers so the penalty cost would be minimized to avoid SLA violation. Another SLA violation-based technique is proposed in the paper (Singh and Viniotis 2016). The author proposed a method to limit the user’s task and schedule it efficiently to minimizes SLA violation. The proposed method divides the user task into multiple subtasks and increases the server’s capacity to decrease the total measurement time of task execution to reduce SLA violation and maximize providers profits. Another SLA-aware resource allocation method for IoT devices is proposed in (Alsaffar et al. 2016), the author presented an architectural based service delegation and resource allocation method for Cloud and fog computing based IoT environment. An SLA and QoS oriented algorithm have been proposed that used linearized decision tree to manage user tasks. A new SLA oriented IoT resource allocation problem handled in (Singh and Viniotis 2017), by buffering scheduling and limit the rate of user’s tasks to achieve SLA. The proposed method conforms

4  Resource Management Techniques for Cloud-Based IoT Environment

75

Table 4.2  SLA-aware IoT resource allocation techniques Algorithm SLA-Based Resource Allocation (Choi and Lim 2016) Cloud-based SLA-aware Resource Allocation (Singh and Viniotis 2016) IoT service delegation and resource allocation (Alsaffar et al. 2016) Resource allocation for IoT applications (Singh and Viniotis 2017)

Improvement Reduce SLA violation and increase system performance

Limitations The proposed method is not compared with other techniques

User’s task is divided into subtasks to reduce SLA violation

The task arrival time is not calculated dynamically

Improve system performance and There is no evidence of practical efficiency implementation. Improve system performance and reduce SLA violation by reducing the measurement time of task arrival.

The proposed method is not working in a multi-tenant environment like multiple data centres

the SLA violation without knowing the arrival pattern of the user’s task in advance and works better in huge traffic of IoT devices in the IoT network. These algorithms and their improvement and limitations are described in Table 4.2.

4.4.2  Context-Aware IoT Resource Allocation Game theory has been applied in many types of research for resource allocation in the device to device communication for IoT devices. Resource allocation is an important aspect for the high performance of data transportation in a wireless network for device to device communication. A location-aware method has been extended in (Huang et al. 2015), that applied the Nash Equilibrium game model for D2D communication in a cellular network. The author proposed a context-aware algorithm that determines the bandwidth of the network to maximize the total use of each station for communication according to the different situations. A cell association problem has been formulated in (Hamidouche et al. 2017) as a two-way matching process between the communicating devices. The proposed model used a correlation among the devices that are present in each area network to enhance the cell association technique and guaranteed to a prevalent outcome. Another one-­ many devices’ association based optimal resource allocation method proposed in (Abedin et  al. 2015) that increased the resource utilization among the network devices. Dynamic QoS requirements have been achieved by the proposed context-­ aware resource allocation method between peer to peer IoT network system. Table 4.3 shows the improvements and limitations of context-aware resource allocation techniques.

76

S. A. Ali et al.

Table 4.3  Context-aware IoT resource allocation techniques Algorithm A game-theory based D2D communication in a Cloud-­ Centric IoT network (Huang et al. 2015) Correlation-based resource allocation method (Hamidouche et al. 2017) Optimal resource allocation method (Abedin et al. 2015)

Improvements D2D communication improved by applied Nash equilibrium game model, maximize bandwidth utilization Repeated information generated by the different devices has been reduced for better performance in cell association method Resource utilization and performance have been increased

Limitations This approach is only Cloud-centric and compared with the existing algorithms The practical evidence is not mentioned in the method The method has not practical evidence

4.4.3  QoS-Aware IoT Resource Allocation Quality of Service (QoS) is an important aspect of any service-based application. The quality of service should be matched with the service level agreement (SLA). Several pieces of research have been done in this area for a different scenario. A QoS-based IoT resource allocation method has been proposed in (Chen et  al. 2012), which reduced the intrusion between the device to device communications. With the help of PFR method and intrusion limited area control method, the use of resources can be restricted to D2D users. The D2D users get the resources according to the wireless network channel gain, which balances the system workload and enhances the system performance. An analytical model for heterogeneous traffic of M2M devices has been proposed in (Shorgin et al. 2015), which used fixed transmission nodes for M2M devices according to the user’s requirements to achieve QoS constraints. An optimization protocol has been proposed in (Colistra et al. 2014a), which is based on the consensus algorithm for robust and efficient resource allocation in the heterogeneous IoT networks. The algorithm considered the task frequency and buffer occupancy of nodes that are involved in the communication. The method is adaptive in dynamic and heterogeneous feature of IoT device networks. Two communication-based resource allocation method for IoT has been proposed in (Colistra et al. 2014b), which used broadcast and gossip methods among the nodes for exchange and update the communication information. The proposed method has been evaluated in three different scenarios: the entire network, single task — single frequency and single task — total frequency. The output error of the system has been reduced to 5% with compared to the centralized solution for the reduction of message transfer and increase the system reliability. QoS-based resource allocation methods with features and limitations are described in Table 4.4.

4  Resource Management Techniques for Cloud-Based IoT Environment

77

Table 4.4  QoS-aware IoT resource allocation techniques Algorithm Downlink resource allocation method (Chen et al. 2012) Radio Resource Allocation Scheme (Shorgin et al. 2015) Task allocation in group of IoT devices (Colistra et al. 2014a) Consensus-based task allocation in IoT (Colistra et al. 2014b)

Improvements The intrusion among the communication channel has been reduced and enhanced the system performance The traffic among the M2M devices has been reduced

Limitations Practical implementation is missing

Gossip and broadcast methods are used in which broadcast gives better results

The QoS and real conditions have less considered.

The method is not implemented in real conditions The optimal resource allocation has been The QoS parameter achieved with 5% error needs more attention

4.4.4  Energy-Aware IoT Resource Allocation Due to the huge amount of heterogeneous and more electric power consumed smart devices, it is important to handle the efficient use of electricity or power to reduce the carbon footprints and produce a green computing environment. In many fields like Cloud Computing, fog computing and edge computing, many researchers have been already done a lot of work to reduce power consumption. In IoT based environment there is also some evidence of work in the reduction of power consumption. In (Baccarelli et al. 2017) a novel resource allocation method for the fog of everything (FoE) has been proposed that presents the energy-delay performance of virtual fog of everything (V-FOE) and compared with the V-D2D technological platform. Three ways of communication between the devices in fog computing architecture network (FOCAN) (Naranjo et al. 2017) has been described to meet the QoS and reduce power usage in an effective way. This approach is used in Fog-­supported smart cities architecture to share infrastructure resources among smart devices. A generalized Nash equilibrium (GNE) approach and its unique conditions are derived in (Abuzainab et al. 2017), to handle the heterogeneity of IoT resources, in terms of QoS and resource constraints. A cognitive hierarchy game theory has been used in this method to enable the devices to reach CH equilibrium (CHE) for rationally corresponds to the heterogeneous computing capability and access information of each MTDs and HTDs. The proposed model reduces energy consumption by MTDs by 78%. Another novel ECIoT architecture has been proposed in (Li et al. 2018), to additional, improve the system performance by control the process admission and control the power consumption of the resource of the IoT system. A Lyapunov stochastic optimization based cross-layer dynamic network optimization method is used in ECIoT to enhance the system utility. Table 4.5 shows the limitations and improvements done in these proposed algorithms.

78

S. A. Ali et al.

Table 4.5  Energy-aware IoT resource allocation techniques Algorithm Energy-efficient resource allocation for Fog of Everything (Baccarelli et al. 2017) FOCAN: A Smart city architecture for resource allocation (Naranjo et al. 2017) Cognitive hierarchy theory for resource allocation (Abuzainab et al. 2017) Joint admission control resource allocation (Li et al. 2018)

Improvements Reduce energy consumption and delay and enhance the performance of the FOE system The method reduces energy consumption and improves latency.

Limitations Real-time testing has not done Implementation is not done

The method reduces power consumption by 78%

The only Simulation is done

Increased System Throughput and reduce the delay between devices communication

No evidence of practical implementation.

4.4.5  Cost-Aware IoT Resource Allocation IoT network consists of many heterogeneous and many powerful resources needed smart devices which are managed by Cloud Computing, fog computing as well as edge computing infrastructures. These multiple network devices of IoT request the resources for their task completion to fulfil the QoS. The total utilization cost has been calculated according to the served resources and activation cost of each interface devices in the network to serve demand. This cost estimation problem is known Service-to-Interface Assignment cost problem. Two SLA methods are proposed mathematically in (Angelakis et al. 2016) to reduce the computational cost. In the first method, the demand for the resource has been fulfilled in one round but in the second method, demand has been fulfilled in multiple rounds. The proposed method splits the activation cost and distributes it in multiple interfaces to reduce the cost of activation. The effective and efficient allocation and release of Cloud resource are important to reduce the service cost of Cloud resource. A multi-agent-based Cloud resource allocation method for IoT devices has been proposed in (Manate et  al. 2015), to audit the dynamic use of resources by the IoT devices. The audit of resource usage help to control the bad resource utilization and increase the u­ tilization of the resources that in turn improve the performance and reduce the total cost of the system. Another method (Lan et al. 2013), based on Stackelberg game model has been proposed to reduce the network resource cost. The method analysis the lower and upper layer of the network and verifies the Nash equilibrium point of the no n-cooperative game between the upper layer of the network to reduce the cost. An iterative method is used to reach the Nash equilibrium by creating the Stackelberg game of the entire network. Multiple heterogeneous network interfaces have been used IoT devices that used a huge amount of services. A resource to heterogeneous service model has been proposed in (Angelakis et  al. 2015), which is based on mixed-integer linear program (MILP) formulation. The cost of services can be reduced by splitting the services over the different interfaces dynamically.

4  Resource Management Techniques for Cloud-Based IoT Environment

79

Table 4.6  Cost-aware IoT resource allocation techniques Algorithm Heterogeneous resource allocation for flexible services (Angelakis et al. 2016) Optimizing cloud resource allocation architecture (Manate et al. 2015) Heterogeneous-oriented Resource Allocation Method (Lan et al. 2013) Flexible Allocation of Heterogeneous Resources (Angelakis et al. 2015)

Improvements The Cost has been reducing by splitting the services among the interfaces Increase system performance, reduce the total cost by minimizing the use of VMs Cost resources usage has been reduced

Limitations The cost has been slightly increased in multi-round method The cost of ideal resource usage must be reduced No evidence of practical implementation

The cost has been reduced and increase the system performance

The more theoretical discussion presents not practically implemented

Improvement and limitations of cost-aware resource allocation techniques for IoT environment have been discussed in Table 4.6.

4.5  Parameters of IoT Resource Management Techniques There are various IoT resource management technique parameters, which are described below. This section aims to answer the following questions related to parameters of resource allocation techniques for IoT.

4.5.1  W  hat Are the Various Parameters of Resource Allocation? Resource allocation is an important aspect of Cloud-based IoT environment. The devices are connected in the IoT environment by the internet and produced a huge amount of data which are stored in Cloud for further analysis to infer useful ­information for the organization and application of the system. Many resource allocation parameters have been discussed in (Ali et al. 2019a, b; Ali and Alam 2016). In this section, the author defines the various parameters of resource allocation for IoT environment which should be considered for improvement in the development of the IoT resource allocation techniques. • Performance It is an amount of work done by the IoT layer to accomplish the on-demand services of the user’s task. Performance of the system should be high. • Throughput The total number of task or job completed by the IoT system is measured as the throughput of the IoT system. Throughput must be high in IoT systems to accomplish all the user’s task.

80

S. A. Ali et al.

• QoS Quality of Service (QoS) is the measurement of the quality services provided by the Cloud system to the end-user which are agreed on the SLA agreement. • Delay Amount of time to respond to a waiting user’s task when the system is busy in another task execution. The delay should be minimum in the channel to enhance the system performance. • Bit Rate It an amount of rate at which IoT devices transfer data from one location to another location. Data transfer rate must be high in internet-based IoT environment. • Reliability It is the ability to execute the given task in time without affecting by the system failure. Reliability of the system should be high to ensure the user’s task completion in time. • SLA Service level agreement between the service provider and the service consumers must be met to overcome the cost overhead and reduce the risk of users and providers relationships violation. • Time It is an important factor in the IoT environment because smart devices produced a huge amount of data regularly. It is a plan to schedule a task in the IoT environment for their execution. • Cost IoT environment uses many services from the Cloud infrastructure providers and in return give money to the service provider. The Cost of services is measured in the performance of the system and how much the system is productive. The cost of the system should be minimum. • Energy IoT environment consists of many network nodes and smart devices which are connected to large data centres. These devices and data centres consumed a huge amount of energy for their proper functioning. The energy consumption should be minimized to reduce cost and make environment carbon footprint-free. • Availability Resource availability is the measure of time and reliability of the resource in the given period. The availability of the resources should be high to reduce the delay in service. • Utilization Resource utilization is the amount of portion, a resource must be occupied by the system. The resources should be efficiently utilized to enhance the system performance and the reliability of the system.

4.5.2  H  ow Much Degree of Improvement Have Been Done in These Parameters? In this paper, the author studied various Cloud-based IoT resource allocation techniques and classified them into the number of groups. In this section, the parameters of resource allocation which are improved by the given techniques in the classification have been discussed and in Table  4.7 these parameters are marked corresponding to the method or algorithm under study. From Fig.  4.7, we can observe that how much a parameter has been improved and which ones require more attention from the research community.

Reference Choi and Lim (2016) Singh and Viniotis (2016) Alsaffar et al. (2016) Singh and Viniotis (2017) Huang et al. (2015) Hamidouche et al. (2017) Abedin et al. (2015) Chen et al. (2012) Shorgin et al. (2015) Colistra et al. (2014a) Colistra et al. (2014b) Baccarelli et al. (2017) Naranjo et al. (2017) Abuzainab et al. (2017) Li et al. (2018) Angelakis et al. (2016) Manate et al. (2015) Lan et al. (2013) Angelakis et al. (2015) – – – √ – – – √ – – – √ – – – – –

√ –

√ √

– √ √ √ √ √

√ √

√ √

√ √ √

– –

– –

– √ – – – √

– –

– –

– – –

Bit Delay Performance rate – √ – – √ –

Table 4.7  IoT resource allocation parameters

– – –

– –

– –

– – √ – – –

– –

– –

– – –

– –

– –

– – – √ – –

– –

– –

– – √

– √

– – –

– –

– –

√ – – – – –

√ √ √ – √ – – –

√ –

– –

– –

– –

– – –

– –

– –

√ – – – – –

√ √

– –

– – –

– –

– –

– – – – – –

– –

√ √

– – –

– –

– –

– – – – – –

– –

– √

– – –

– –

– –

– – – – – –

– –

– –

√ √ √

– √

– –

– – – – – –

– –

– –

– – –

√ –

√ √

– – – – – √

– –

– –

– – –

– –

√ –

– – – – – –

– –

– –

Context-­ Reliability Throughput QoS Utilization Aware SLA Time Availability Cost Power Latency – – – – – √ – – – – – √ – – – – √ √ √ – – –

4  Resource Management Techniques for Cloud-Based IoT Environment 81

82

S. A. Ali et al.

14 12 10 8 6 4 2 0

SLA-Aware Context-Aware

QoS-Aware

Power-Aware

Cost-Aware

Performance

Bit rate

Reliability

Throughput

QoS

Utilization

Context-Aware

SLA

Time

Availiblity

Cost

Power

Latency

Delay

Fig. 4.7  Degree of improvement in resource allocation parameters

Number of Algorithms

18 16 14 12 10 8 6 4 2

Fig. 4.8  Resource allocation parameters and the corresponding number of algorithms

y nc

te

we r

Resource Allocation Parameters

La

st

Po

Co

lity

e

Av a

ilib

A

Ti m

SL

io n -A wa re

Co n

te

xt

at

oS

Ut

Re l

iliz

Q

tr at e ia bi lity Th ro ug hp ut

ce

Bi

an

rm

Pe

rfo

De

la

y

0

4  Resource Management Techniques for Cloud-Based IoT Environment

83

4.5.3  H  ow Much Improvement Is Required in the Remaining Parameters? In the above section, we have already seen that the number of resource allocation parameters have been improved in the proposed algorithms or methods. But still, various resource allocation metrics need improvement and require much attention from the researchers. In Fig. 4.8, various parameters and their corresponding number of algorithms in which they are improved have been shown. From the figure, we can easily identify that the performance of the system has been much improved by many researchers. But the reliability, latency, availability of resources and delay in communication channel have not been improved so much. Therefore, there is a scope for the researchers to propose more resource allocation algorithms that can improve more and more these resource allocation parameters. Multi-objective and nature-inspired algorithms can be implemented by the researchers.

4.6  Challenges and Issues The IoT environment provides potential services to improve the productivity of the Cloud-based IoT ecosystem. Regardless of its potential benefits, there are silently many challenges are present in the Cloud-based IoT resource allocation. In this paper, there are many research papers studied related to IoT resource allocation problem. Most of the papers are based on one or two aspects of the resource allocation, there should be more work needed in the collaboration of IoT with Cloud, Fog and Edge computing for better implementation of IoT ecosystem. Most of the research work done so far has based on simulation and no evidence of implementation of the proposed algorithm for real IoT ecosystem. Real IoT ecosystem has its own challenges and properties and when these algorithms are implemented in real IoT world then there may be some problems occur because of no proper implementation has been tested in real IoT environment. Another challenge in the resource allocation methods studied in this paper is that all aspect of resource allocation like resource discovery, resource modelling, resource provisioning, resource scheduling, resource estimation and resource monitoring has not considered. Optimization in resource allocation techniques is another issue, more of the research only work on the resource allocation technique but optimization in resource allocation is more important for effective and efficient IoT environment.

84

S. A. Ali et al.

4.7  Future Directions IoT environment makes the life of human being so easy. All daily routine work can be done efficiently with the help of IoT devices. Many industries are planning to launch more IoT devices that can make an easier daily task for the human. Many other technologies give life to the IoT infrastructure and provide services. Cloud-­ based IoT environment for industries application has a future for implementation. The reliability, productivity and cost-effective industry-based applications require more effective and efficient resource allocation. These applications are known as Industries Internet of Things (IIoT) application. Data captured by the smart IoT devices are stored in Cloud and further processed in Cloud infrastructure for inferring knowledge. The amount of data captured by IoT devices is very large that why many researchers proposed Big data as a service to the other industries and Cloud users (Khan et al. 2018, 2019). Machine learning-based resource allocation has a scope to automate the resource allocation in IoT infrastructure. ML-based resource allocation techniques can overcome the scarcity, over and underutilization of Cloud resources. Blockchain technology is also an emerging technology which is used for cryptocurrency (Dittmann and Jelitto 2019). Blockchain is a secure method for transferring data from one node to another. It provides privacy and secure communication in a P2P network. Many types of research have been done in IoT and blockchain integration with Cloud Computing (Abedin et al. 2015; Qiu et al. n.d.), but the blockchain is not used in Cloud-based IoT resource allocation.

4.8  Conclusion Resource allocation is an important aspect of Cloud-based IoT environment for effective and efficient working of the novel paradigm. In this paper, Cloud-based resource allocation techniques for IoT environment have been studied and categorized into different groups: SLA-Aware, Context-Aware, QoS-Aware, Power-Aware and Cost-Aware resource allocation. The author systematically reviewed all these techniques and find out the limitations and improvements of these algorithms. Moreover, resource allocation parameters determined for improvements have also listed and defined. A graphical representation of the degree of improvement in these parameters has been shown in the paper. Lastly, challenges and future directions of the study have been discussed.

References Abedin, S. F., Alam, M. G. R., Il, S., & Moon, C. S. H. (2015). An optimal resource allocation scheme for Fog based P2P IoT Network. In: 년 동계학술발표회 논문집 (pp 395–397). Abuzainab, N., Saad, W., Hong, C. S., &Poor, H. V. (2017). Cognitive hierarchy theory for distributed resource allocation in the Internet of Things, arXiv preprint.

4  Resource Management Techniques for Cloud-Based IoT Environment

85

Alaba, F. A., Othman, M., Hashem, I. A. T., & Alotaibi, F. (2017). Internet of Things security: A survey. Journal of Network and Computer Applications, 88, 10–28. Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., & Ayyash, M. (2015). Internet of things: A survey on enabling technologies, protocols, and applications. IEEE Communications Surveys & Tutorials, 17, 2347–2376. Ali, S. A., & Alam, M. (2016). A relative study of task scheduling algorithms in cloud computing environment. In 2016 2nd International Conference on Contemporary Computing and In- formatics (IC3I) (pp. 105–111). https://doi.org/10.1109/IC3I.2016.7917943. Ali, S. A., Affan, M., & Alam, M. (2019a). A study of efficient energy management techniques for cloud computing environment. In 2019 9th international conference on cloud computing (pp. 13–18). https://doi.org/10.1109/CONFLUENCE.2019.8776977. Ali, S.  A., Khan, S., & Alam, M. (2019b). Resource-aware min-min (RAMM) algorithm for resource allocation in cloud computing environment. International Journal of Recent Technology and Engineering, 8(3), 1863–1870. https://doi.org/10.35940/ijrte.C5197.098319. Alsaffar, A. A., Pham, H. P., Hong, C. S., Huh, E. N., & Aazam, M. (2016). An architecture of IoT ser- vice delegation and resource allocation based on collaboration between fog and cloud computing. Mobile Information Systems. https://doi.org/10.1155/2016/6123234. Angelakis, V., Avgouleas, I., Pappas, N., & Yuan, D. (2015). Flexible allocation of heterogeneous re- sources to services on an IoT device. In 2015 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS) (pp. 99–100). Angelakis, V., Avgouleas, I., Pappas, N., Fitzgerald, E., & Yuan, D. (2016). Allocation of heteroge- neous resources of an IoT device to flexible services. IEEE Internet of Things Journal, 3, 691–700. Atlam, H. F., Alenezi, A., Walters, R. J., Wills, G. B., & Daniel, J. (2017). Developing an adaptive risk- based access control model for the Internet of Things. In 2017 IEEE International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (Green- Com) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData) (pp. 655–661). Atzori, L., Iera, A., & Morabito, G. (2010). The internet of things: A survey. Computer Networks, 54(15), 2787–2805. Baccarelli, E., Naranjo, P. G. V., Scarpiniti, M., Shojafar, M., & Abawajy, J. H. (2017). Fog of every- thing: Energy-efficient networked computing architectures, research challenges, and a case study. IEEE Access, 5, 9882–9910. Baccelli, E., Hahm, O., Günes, M., Wählisch, M., & Schmidt, T. C. (2013). RIOT OS: Towards an OS for the Internet of Things. In Proceedings of the IEEE conference INFOCOM WKSHPS (pp. 79–80). Bassi, A., Bauer, M., Fiedler, M., & Kranenburg, R. V. (2013). Enabling things to talk. New York: Springer. Cao, Q., Abdelzaher, T., Stankovic, J., & He, T. (2008). The liteos operating system: Towards unix-­ like abstractions for wireless sensor networks. In Proceedings of the international conference on information processing in sensor networks (pp. 233–244). Chen, X., Chen, L., Zeng, M., Zhang, X., & Yang, D. (2012). Downlink resource allocation for device-to-device communication underlying cellular networks. In 2012 IEEE 23rd international symposium on Personal Indoor and Mobile Radio Communications (PIMRC) (pp. 232–237). Choi, Y., & Lim, Y. (2016). Optimization approach for resource allocation on cloud computing for IoT. Journal of Distributed Sensor Networks. https://doi.org/10.1155/2016/3479247. Colistra, G., Pilloni, V., & Atzori, L. (2014a). Task allocation in group of nodes in the IoT: A consensus approach. In 2014 IEEE International Conference on Communications (ICC) (pp. 3848–3853). Colistra, G., Pilloni, V., & Atzori, L. (2014b). The problem of task allocation in the Internet of Things and the consensus-based approach. Computer Networks, 73, 98–111. Crosby, G. V., & Vafa, F. (2013). Wireless sensor networks and LTE-A network convergence. In Proceedings of the IEEE 38th Conference on Local Computer Networks (LCN) (pp. 731–734).

86

S. A. Ali et al.

Delicato, F. C., Pires, P. F., & Batista, T. (2017). The resource management challenge in IoT. In Resource Management for Internet of Things (pp. 7–18). Springer. Dittmann, G., & Jelitto, J. (2019). A Blockchain proxy for lightweight IoT devices. In 2019 Crypto Valley Conference on Blockchain Technology (CVCBT) (pp. 82–85). https://doi.org/10.1109/ CVCBT.2019.00015. Evans, D. (2011). The Internet of Things how the next evolution of the internet is changing everything. Ferro, E., & Potorti, F. (2005). Bluetooth and Wi-fi wireless protocols: A survey and a comparison. IEEE Wireless Communications, 12, 12–26. Gigli, M., & Koo, S. (2011). Internet of things: Services and applications categorization. Advanced Internet of Things, 01, 27–31. Hamidouche, K., Saad, W., & Debbah, M. (2017). Popular matching games for correlation-aware re- source allocation in the internet of things. In IEEE International Symposium on Information Theory (ISIT) submitted to IEEE. Horrow, S., & Sardana, A. (2012). Identity management framework for cloud based internet of things. In Proceedings of the first international conference on Security of Internet of Things (pp. 200–203). Huang, J., Yin, Y., Duan, Q., & Yan, H. (2015). A game-theoretic analysis on context-aware resource allocation for device-to-device communications in cloud-centric internet of things. In 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud) (pp. 80–86). Khan, S., Shakil, K. A., Ali, S., & Alam, M. (2018). On designing a generic framework for big data as-a-service. In 2018 1st International Conference on Advanced Research in Engineering Sciences (ARES) (pp. 1–5). https://doi.org/10.1109/ARESX.2018.8723269. Khan, S., Ali, S. A., Hasan, N., Shakil, K. A., & Alam, M. (2019). Big data scientific workflows in the cloud: Challenges and future prospects. In H. Das, R. K. Barik, H. Dubey, & D. S. Roy (Eds.), Cloud computing for geospatial big data analytics (Vol. 49). Cham: Springer. Koshizuka, N., & Sakamura, K. (2010). Ubiquitous ID: Standards for ubiquitous computing and the Internet of Things. IEEE Pervasive Comput. Sensors, 9, 37. Lan, H.  Y., Song, H.  T., Liu, H.  B., & Zhang, G.  Y. (2013). Heterogeneous oriented resource allocation method in internet of things. Applied Mechanics and Materials, 427, 2791–2794. Lee, I., & Lee, K. (2015). The Internet of Things (IoT): Applications, investments, and challenges for enterprises. Business Horizons, 58, 431–440. Levis, P., Madden, S., Polastre, J., Szewczyk, R., Whitehouse, K., Woo, A., Gay, D., Hill, J., Welsh, M., & Brewer, E. (2005). An operating system for sensor networks. Li, S., Zhang, N., Lin, S., Kong, L., Katangur, A., & Khan, M. K. (2018). Joint admission control and resource allocation in edge computing for Internet of Things. IEEE Network, 32, 72–79. Lin, J., Yu, W., Zhang, N., Yang, X., Zhang, H., & Zhao, W. (2017). A survey on internet of things: Architecture, enabling technologies, security and privacy, and applications. IEEE Internet of Things Journal, 4, 1125–1142. Malik, A., & Om, H. (2018). Cloud computing and internet of things integration: Architecture, applications, issues, and challenges. In Sustainable cloud and energy services (pp.  1–24). Springer. Manate, B., Fortis, T. F., & Negru, V. (2015). Optimizing cloud resources allocation for an Internet of Things architecture. Scalable Comput, 15, 345–355. Marques, G., Garcia, N., & Pombo, N. (2017). A survey on IoT: Architectures, elements, applications, QoS, platforms and security concepts. In Advances in Mobile cloud computing and big data in the 5G era (pp. 115–130). Springer. Mattern, F., & Floerkemeier, C. (2010). From active data management to event-based systems and more. New York: Springer. From the internet of computers to the Internet of Things. Mcdermott-Wells, P. (2004). What is bluetooth? IEEE Potentials, 23, 33–35. Mell, P., & Grance, T. (2011). The NIST definition of cloud. Computing.

4  Resource Management Techniques for Cloud-Based IoT Environment

87

Muntjir, M., Rahul, M., & Alhumyani, H. A. (2017). An analysis of internet of things (IoT): Novel architectures, modern applications, security aspects and future scope with latest case studies. International Journal of Engineering Research and Technology, 6(6), 422–447. Naranjo, P. G. V., Pooranian, Z., Shojafar, M., Conti, M., & Buyya, R. (2017). FOCAN: A fog-­ supported smart city network architecture for management of applications in the internet of everything environments. Journal of Parallel and Distributed Computing, arXiv, preprint. Pourghebleh, B., & Navimipour, N.  J. (2017). Data aggregation mechanisms in the Internet of Things: A systematic review of the literature and recommendations for future research. Journal of Network and Computer Applications, 97, 23–34. Qiu, C., Yao, H., Jiang, C., Guo, S., & Xu, F. (n.d.). Cloud computing assisted blockchainenabled Internet of Things. IEEE Transactions on Cloud Computing. https://doi.org/10.1109/ TCC.2019.2930259. Shorgin, S., Samouylov, K. E., Gaidamaka, Y. V., Chukarin, A., Buturlin, I. A., & Begishev, V. (2015). Modeling radio resource allocation scheme with fixed transmission zones for multiservice M2 M communications in wireless IoT infrastructure. ACIIDS, 2, 473–483. Singh, A., & Viniotis, Y. (2016). An SLA-based resource allocation for IoT applications in cloud environments. Cloudification of the Internet of Things (CIoT), 1–6. Singh, A., & Viniotis, Y. (2017). Resource allocation for IoT applications in cloud environment s. International Conference on Computing, Networking and Communications (ICNC), 719–723. 2017. Stallings, W. (2015). The Internet of Things: Network and security architecture. Internet Protocol Journal, 18(4), 381–385. Subashini, S., & Kavitha, V. (2011). A survey on security issues in service delivery models of cloud computing. Journal of Network and Computer Applications, 34(1), 1–11. Varghese, B., & Buyya, R. (2018). Next generation cloud computing: New trends and research direc- tions. Future Generation Computer Systems, 79, 849–861. Wang, L., Laszewski, G. V., Younge, A., He, X., Kunze, M., Tao, J., & Fu, C. (2010). Cloud computing: A perspective study. New Generation Computing, 28(2), 137–146. Want, R. (2006). An introduction to RFID technology. IEEE Pervasive Computing, 5, 25–33. Want, R. (2011). Near field communication. IEEE Pervasive Computing, 10, 4–7. Xing, X.  J., Wang, J.  L., & Li, M.  D. (2010). Services and key technologies of the Internet of Things. ZTE Commun, 2, 11–11. Yan, Z., Zhang, P., & Vasilakos, A. V. (2014). A survey on trust management for Internet of Things. Journal of Network and Computer Applications, 42, 120–134. Yang, L., Yang, S. H., & Plotnick, L. (2013). How the Internet of Things technology enhances emergency response operations. Technological Forecasting and Social Change, 80, 1854–1867. Zhang, Q., Cheng, L., & Boutaba, R. (2010). Cloud computing: State-of-the-art and research challenges. Journal of Internet Services and Applications, 1(1), 7–18.

Chapter 5

Data Management for the Internet of Things Amrit Sahani, Ranjit Kumar, Suchismita Chinara, Anjali Kumari, and Bina Patro

Abstract  IoT is the administration systems which are worldview used and where all the interconnected, keen articles constantly create information and sends it over the Internet. A significant part of the IoT activities is equipped in the direction of assembling minimal effort and vitality proficient equipment for these articles, just as the correspondence innovations that give objects interconnectivity. Moreover, the answers and solutions for managing and using the monstrous volume of information created by these articles of IoT are yet to develop. Conventional database arrangements miss the mark in fulfilling the refined application demands for an IoT device that has a global scale. The recent answers for IoT information the board conveys halfway parts for IoT conditions along an extraordinary spotlight on sensor systems and interconnection. This chapter provides an overview of information and the arrangements which are opted for IoT and the subsystems of IoT.  We feature the unmistakable and distinctive plan that we accept to be tended into IoT data management solutions and examine how they are drawn closer by the proposed arrangements. We, at last, suggest the information and the board structure for IoT that deliberates over the talked about plan components and goes about as a seed to an exhaustive IoT information arrangement procedure. The framework we work will adopt a federated, data- and the information-significant method to loop the diversified items along with the profusion of information for the devices that are coined for IoT. Keywords  IoT · Sensor · Data management · Framework

A. Sahani (*) Siskha ‘O’ Anusandhan University, Bhubaneswar, Odisha, India Institute of Technical Education and Research, Bhubaneswar, Odisha, India R. Kumar · S. Chinara National Institute of Technology, Rourkela, Odisha, India A. Kumari College of Engineering Roorkee, Roorkee, Uttarakhand, India B. Patro University of Berhampur, Berhampur, Odisha, India © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_5

89

90

A. Sahani et al.

5.1  Introduction The Internet of Things is where each and every physical element is allotted with the web through the channel of framework devices, switches or routers and trade data and exchange information. IoT empowers things to be controlled remotely across over existing framework establishment. The idea of IoT is as a rule incredible and technically knowledgeable framework which reduces human effort similarly as straightforward access to physical contraptions. This procedure has autonomous control including any contraption which gets controlled with no human correspondence (Vermesan et al. 2009). The objective of IoT is to try is to attain and provide a stage for making pleasing organizations, co-employable administrations and applications that outfit the total force of advantages for the individual “Things” and other subsystems planned for handling the recently referenced “Things”. Any point of convergence of such advantages has the plenitude of information which is made available by following the mixing of data and is conveyed continuously similarly as data set away in never-ending storage facilities and archives. This information can make the affirmation of creative and unique applications and worth included organizations possible and will give a significant source to float assessment and essential possibilities. A careful organization arrangement of information which is made, set away from the following articles inside IoT is henceforth expected to attain the referenced target (Pujolle 2006). The board of information is a wide and far reaching idea inferring to the designs, models, usability, and strategies in running the organization properly although the data lifecycle needs of a particular structure. Taking into the thought with regards to IoT the information the executives ought to go as a layer between the items and gadgets producing the information and the applications which are getting to the information for investigation, assessment and organizations. The contraptions are to be organized into different underlying systems with individuality in an organization inside which different levels are available for the boarding. The convenience and data given by these subsystems are to be made available to the course of action of IoT, dependent upon layers of security needed by the owners of the underlying-­ levels (Wu et al. 2011). IoT information ought to have specific traits which can make standard and conventional social based database an obsolete game plan. An enormous volume of heterogeneous, spilling and topographically dissipated progressing data will be made by countless various devices once in a while transmitting signals over particular watched and checked marvels else specifying about the occasion of peculiar interesting events. Time to time supervision and observations taken periodically which are similar to correspondence overhead and limit on account of their spilling and unending quality, whereas, occurrences that are available timely and responsively from the beginning till the finished rebounding reaction times depending upon the intensity irritability of the reply needed to complete the assigned task (Cooper and James 2009).

5  Data Management for the Internet of Things

91

It has been foreseen that there is a reestablished enthusiasm for database structures investigating the spotlights on substitute representation rather than standard conventional prototype. The move from ordinary database models has different viewpoints that are especially profitable to IoT, for instance, the utilization of remote amassing at the Things layer, non-assistant data support, loosening up of the Atomicity, Consistency, Isolation, and Durability (ACID) properties to trade off consistency and availability, and blend of indispensable profitability as a combination of vitality effectiveness just as information the executives plan crude.

5.2  Management of Data in IoT Conventional information and data systems management framework, which handle the capacity, recovery, update the basic information and organize records. In respect to the IoT, information the administrative frameworks must abridge information online and through different perspective while giving stockpiling and inspecting offices, also facilities for auditing for disconnected and offline investigation through analysis (Ramakrishnan and Gehrke 2002; Cattell 2010). This grows the information idea on the board from offline capacity, question (querying) handling, and exchange the operational activities into online-offline correspondence/storage dual tasks. We carry out the settings of IoT and examine the parts of the IoT and after that layout the vitality utilization for different parts and stages accomplished so as to have a superior comprehension of the information management for the Internet of Things.

5.2.1  Life-Cycle of Information The circulated information, and resources inside an IoT framework—continues from information generation to accumulation, move, discretionary sifting, filtering and preprocessing, lastly to the storage, file-logging and archiving. Questioning (Querying) and the investigation is the end focuses that initiate demand and expend information creation, yet information generation may be opted to be “pushed” to the IoT devouring administrations and preparing for the client’s requirement. This includes creation, accumulation, separation, and some essential basic querying and preparatory handling operational tasks, which can be viewed on the web. Concentrated preprocessing, long-term stockpiling information and archiving data are considered for disconnected/offline storage activities (Ozsu and Valduriez 2011). Storage tasks target making information accessible in the extended term for steady access. On the other hand, authentication is worried just about only reading the information. As many of the frameworks of Internet Of Things(IoT) may produce, process, and store information in-organize for ongoing and restricted administrations, with no compelling reason to proliferate this information further up to

92

A. Sahani et al.

focus focuses on the framework, “edges” that consolidate both preparing and capacity components may exist as independent items in the propaganda. In the accompanying sections, every component in the IoT information phenomenon is clarified (Pungila et al. 2009). Questioning Intensive information frameworks depend on questioning (querying) for the centre procedure to get to and recover information. With regards to IoT, an inquiry can be issued either to demand continuous information to be gathered for transient checking purposes or to recover a specific perspective on the information put away inside the system framework. The primary case is regular when a (for the most part confined) constant request for information is required. The subsequent case speaks to more globalized perspectives on information and top to bottom examination of patterns and examples (Sen and Ramamritham 2005). Creation (Production) Data generation includes detecting and move of information by the “Things” inside the IoT system and announcing this information to invested individuals at irregular intervals (as in a buy-in/sell model), pushing it up the system to conglomeration focuses and along these lines for the servers of database, or transmitting it through a reaction activated from inquiries which seek the information through the devices which act as a sensor and keen determined articles. Information is normally time-­ stepped, conceivably geo-stepped and globalized where it can be as straightforward key-esteem sets, or it might contain superior quality sound/picture/visuals, along with differing ranks of multifaceted nature of varying natures of complexity (Güting and Mamoulis 2011). Accumulation (Heap) The sensing materials and keen items inside the IoT accumulate the information in a specific time interim and account it to administer parts. Information might be gathered at focus focuses or entryways inside the system where it is additionally sifted and prepared, and potentially intertwined into reduced structures for productive transmission. Remote correspondence advancements, for example, Wi-Fi cells being utilized by articles to send information for gathering focuses. Total (Aggregation) Remitting out all the crude information from the systems progressively is restrictively costly given the expanding information gushing rates and the constrained transfer speed. Collection and combination strategies send synopsis and consolidating activities progressively to pack the volume of information to be put away and transmitted (Spiess et al. 2009). Conveyance As information is separated, amassed, collected and perhaps prepared from the fixation focuses else through the independent indirect elements inside the IoT, consequences from the particular procedures might be forbade for further framework, in a way as definite reactions, or for capacity and top to the bottom investigation.

5  Data Management for the Internet of Things

93

Wired or remote broadband correspondences might be utilized there to move information to last information at stable, secured and established repositories (Guinard et al. 2010). Preliminary Processing The information through the IoT might originate through various provenances in accordance to fluctuating arrangements and patterns. Information may be preprocessed to deal with missing information, expel redundancies and incorporate information from various sources into a brought together outline before being focused on storage capacity. The preliminary processing, a known system in information mining, is called information cleansing. Pattern combination do not infer forceful power fitting of the considerable number of information into a fixed social (tables) composition, yet rather an increasingly conceptual meaning of a steady method to get to the information without tweaking approach in the information format(s). Randomness at various steps in the mapping might be added at this stage to IoT information things so as to deal with a vulnerability that might be available in the information or to manage the absence of trust which might have existence in information repositories. Capacity/Competency The following stage controls the productive stockpiling association of information similar to the consistent reconditioning of information along with fresh data as it winds up accessible. Filing alludes to the disconnected long haul stockpiling of information that isn’t quickly required for the framework’s continuous activities. The centre brought together capacity is the sending of capacity structures that adjust to the different information types and the recurrence of information catch. Social database the board frameworks are a well-known decision that includes the association of information into a table outline with predefined interrelationships and metadata for proficient recovery at later NoSQL key-esteem stores are picking up prevalence as capacity advancements for their help of enormous information stockpiling with no dependence on social composition or solid consistency prerequisites run of the mill of social database frameworks. Capacity can likewise be decentralized for self-sufficient IoT frameworks, where information is kept at the articles that produce it and isn’t sent up the framework. Be that as it may, because of the constrained abilities of such articles, stockpiling limit stays restricted in contrast with the unified stockpiling model. Handling/Analysis This stage includes the continuous recovery and examination activities performed and put away and chronicled-archived information so as to pick up bits of knowledge into recorded information and foresee future patterns, or to recognize variations from the norm out from the information which may trigger for the upcoming examination or activity. Undertaking explicit pretreatment might be expected to channel and clean information before important tasks happen. At the point when a subsystem of IoT is self-sufficient and need not require a perpetual capacity of its information, rather keeps the handling and capacity for the following system, at that

94

A. Sahani et al.

point in-organize preparing might be performed in light of ongoing or limited inquiries. Glancing back, the progression of information may take one of three ways: a way for independent frameworks inside the IoT that returns from inquiry to generation to in-arrange handling and after that conveyance, a way that begins from creation and continues to accumulation and sifting/conglomeration/combination and finishes with information conveyance to starting (potentially worldwide or close ongoing) inquiries, lastly a way that stretches out the generation to total further and incorporates preprocessing, perpetual information stockpiling and authentic, and top to bottom preparing and investigation. In the following area, the requirement for information the executive’s arrangements that outperform the present capacities of customary information the board is featured in light of the recently sketched out life cycle.

5.2.2  M  anagement of Information for IoT and Data Management Systems for the Traditional Methods Following the view of the IoT information lifecycle examined thoroughly, we can separate the working of IoT information executive’s framework into a real-time web application that can continuously communicate straightforwardly with the any other interconnected IoT articles and sensors, and with the help of a offline backend system that can handles the mass stock-piling, gathering of the information and examination from top to bottom of IoT information. The information management frontend is correspondence escalated; including the spread of inquiry demands and result from the sensors and tech-savvy materials. The backend is capacity concentrated; including the data stockpiling of created information for the preparation of later and investigation for more top to bottom questions. In spite of the fact that the capacity components live toward the back, they communicate with the frontend on a successive premise by means of constant updates and are therefore implied on the web. The self-ruling autonomous limits in the existence process which is viewed as more correspondence concentrated than storage escalated, as they give constant information to specific inquiries (Sen and Ramamritham 2005). This imagined information and the board engineering parts significantly oozing out from current management frameworks of database, which are primarily working for storage. The conventional frameworks, which have the main part of the information gathered from predefined and limited sources and put away in scalar structure as indicated by severe standardization controls in relations. Inquiries are utilized to recover explicit “rundown” perspectives on the framework or update explicit things in the database. New information is embedded in the database when required, likewise by means of addition inquiries. Inquiry activities are normally neighborhood, with execution costs bound to handling and middle of the road stockpiling. Exchange the board instruments ensure the ACID properties so as to authorize generally

5  Data Management for the Internet of Things

95

speaking information uprightness. Regardless of whether the database is appropriated over various locales, inquiry preparing and conveyed exchange the board are implemented. The execution of appropriated questions depends on the straightforwardness guideline, which manages that the database is still seen intelligently as one unified unit, and the ACID properties are ensured through the two-stage submit convention (Cattell 2010). The frameworks of IoT, imagination is drastically unique, along with a gigantic—and development of information sources; sensors, RFIDs, inserted frameworks, and portable devices. In spite of infrequent updates and inquiries submitted to customary Data Base Management Systems (DBMS) information is spilling always from the large number of “Things” to IoT for storing information, and questions increasingly visit and with progressively adaptable needs. Various leveled information detailing and collection might be required for versatility ensure just as to empower progressively brief handling usefulness. The severe social database pattern and the social standardization practice might be loose for progressively unstructured and adaptable structures that adjust to the various information types and refined questions. Albeit dispersed Database Management Systems (DBMSs) advance inquiries dependent on correspondence contemplations, streamlining agents base their choices on fixed and well-­ characterized mappings. This may not be the situation in IoT, where new information sources and spilling, confined information make an exceptionally unique condition for inquiry streamlining agents. Endeavoring to ensure the straightforwardness prerequisites forced in appropriated DBMSs on IoT information the board frameworks is testing, if certainly feasible. Moreover, straightforwardness may not be required in IoT, on the grounds that imaginative applications and administrations may require area and setting mindfulness. Keeping up the mind of ACID properties in limited subsystems when the changes are being executed and exchanges will be overseen, yet trying for more space in a globalized way. Nonetheless, the component of versatile information resources and singly created information may be joined into the formally settled information extent is a novel test which is to be tended from the devices of IoT information and management frameworks (Hauer et al. 2008). The most accompanying area gives subtleties of certain information that the board answers for IoT that incorporates most of the things above in an operational mechanism. Those propositions which then dissected a lot of structure natives which is regarded essential for the information management for IoT are seen.

96

A. Sahani et al.

5.3  S  ystematic Survey of the Management for the Devices of IoT and Design Primitives Numerous different kinds of elementary designs are possible that can be used to regulate the logical, insightful, tangible and physical structure for the management for the IoT solutions. Identifying the basic elementary designs is important in constructing a comprehensive solution for the data administration. These designs categorize into the three major segments: Collection of Data System design for Data management Refining and Processing of the Data Collection of data elements make the discovery of the objective and objects of IoT and sub-systems identifiable where data is given to data repositories. This includes management of data and design for the architectural components of the data management system and the process in which data is being stored and managed in the design framework. Finally, the processing of the data and information is done with the real-time access to actual data (Cooper and James 2009).

5.3.1  Data Collection and Information Management Systems One of the worth included administrations that IoT is foreseen is to give is the capacity to take advantage of assorted information that might not really have a place with the equivalent proportional IoT subsystem. Along these lines, at the origin recovery implementation is required for applications using IoT with the goal of reporting their administration requirements and getting reactions from sources whose information might fulfill – those necessities. On the other hand, sources can intermittently declare their administrations; the information they can report and create. A model structure that tends to the disclosure of information sources as a fundamental piece of information the board is proposed in. A model structure that keeps an eye on the exposure of data sources as a crucial snippet of data the will be proposed. The framework on which the system is established may find out the data sources either by methods for creeping, or by methods having pre-defined sources of data that can build new ones as found. Essentialness gainful responses for managing data from compact sources of data that were outlined. Desire of position has been cut down to imperativeness usage of position following advances, and setting or checking was proposed to cut down the pointless counts and correspondence. Position expectation was utilized to bring down the vitality utilization of position following advances, and setting checking

5  Data Management for the Internet of Things

97

was proposed to diminish superfluous computational calculations and correspondence (Ozsu and Valduriez 2011). 5.3.1.1  Collection Strategy for Data The accumulation of information from the “Internet of Things” layer might be temporal or modular. Fleeting information gathering incorporates gathering information from all “Things” at determined interims, or stipulated time intervals while modular accumulation includes gathering information relating to explicit components. The variety assortment of information needs will be normal in IoT (Internet of Things) frameworks which might oversee in having more than the required database example for obliging the dual information aggregation approaches (Pungila et al. 2009).

5.3.2  Design Elements for Database Framework United Architectural Planning Theories from conveyed, unified data directory frameworks have to be adjusted for the requirements of IoT (Internet Of Things) information administration. Disseminated database frameworks deal with a solitary and single database conveyed over numerous locales. Federated database frameworks, then again, oversee autonomous and conceivably heterogeneous information stores at different locales. In a unified design, at the complete and the full self-sufficiency along the information is kept up when it takes part in a repository alliance, in light of that site’s operational necessities. On the off chance that the various subsystems shaping IoT are to be seen as free “information manufacturing plants” with self-governing database frameworks, at that point it is fascinating to investigate the adjustment of inquiry optimization focused at circulated heterogeneous database frameworks to the characteristically disseminated and heterogeneous IoT subsystems (Pujolle 2006). Information and Sources-Driven IoT Middleware The requirement in information-driven software acts as a gap between the correspondence-­driven “things” and the storage-driven information repositories. This middleware layer will serve to give increasingly adaptable abilities to finding information sources and getting to and dissecting heterogeneous and topographically appropriated information, data distributed throughout. The intermediate layer will likewise focus on adaptable exchange in huge chunks of information volume sent from the system and is stored in the information repositories. The need for using maximum capacity of information is to be tackled in various information sources inside IoT by giving the way to find such useful reference which have

98

A. Sahani et al.

dependence with methodology, area or worldwide question needs that are skeptic for the fundamental system (Agarwal and Alam 2018). An engineering which joins the middle software for information handling in enormous scale WSNs to cover arrange for diversity and encourage information conglomeration that are proposed for. The center and main component in the mechanism is a virtual sensing device which coordinates different information streams from genuine sensors in bottom layers and to a solitary information branches. Information branch inquiries has to be assessed, the outcomes from the devices are to be put away in brief relations, and the consequence of joining these branches is put away forever just whenever required, else needed by the application (Pujolle 2006; Wu et al. 2011). Adaptable Prototype The database using relational model frameworks, standards of individuality at, practical reliance, dependency on functional activities and standardization utilized for the characterization of the framework of the relations in the model of the database structure. The models of database which can leave from the social model for an increasingly adaptable database structure are now-a-days increasing extensively and found everywhere, despite the fact that it’s been demonstrated that social and analogous DBMSs knocks out the unstructured ideal models and is way ahead of the traditional data base model. Layered Capacity Stage Contrary to conventional database frameworks, where “steady” information is gathered from the earlier and put away in incorporated, appropriated storage networks, Information from IoT gets constantly refreshed as the level of things gets an observation phenomenal substitute. Regardless of whether it was feasible for the majority of information produced by the Internet of things in any one end so as to be put away in changeless permanent capacity, it will rapidly end up old and new information will wind up accessible. Stream and ongoing databases are not new ideal models and their standards were utilized for natural sensor systems. Be that as it may, the structure of such frameworks tends to concentrated capacity areas and isn’t appropriate in the scaling and conveyance in the maintenance of information capacity created by the devices using IoT, indeed requiring much proficient update and accessing mechanism process (Agarwal and Alam 2018). Two methodologies are to be received for addressing the capacity issue areas of IoT information: every frameworks, which make out to utilize the IoT is its solely devoted repository framework, putting away information from the units in the memory and articles that produce the information and treat it as a database having decentralized. Through the principle perspective, crude or mostly collected information is moved to total focuses and mass storerooms inside an IoT subsystem. The most committed storerooms required to various frameworks shaping IoT (Internet of Things) represents various difficulties. Database access issues, for example, question/exchange preparing and simultaneousness control should embrace the proposed mechanism to find IoT frameworks with storage since it isn’t plausible to give a limited rundown of the considerable

5  Data Management for the Internet of Things

99

number of frameworks connected to the IoT foundation. The presence of exclusive IoT frameworks will direct a requirement for qualifications the executives and consistent access authorizations. The substantial reliance on the transmission of spilling information from sources to storerooms will bring about high correspondence costs. This is made progressively entangled by the requirement for various correspondence advancements for various item classes and transmission situations (Chen et al. 2010; Ramakrishnan and Gehrke 2002).

5.3.3  Preparing Elements 5.3.3.1  Access Model In request to get to information, questioning dialects have been utilized for social frameworks, and later adjusted to sensing systems. SQL (Structured Query Language) provides the true calibration for accessing information, along with leveled choice/join/prediction/projection/total operational mechanism which can be settled in complicated complex questions. Extra builds is being added to the language for compelling new and different categories of information, for example, TinySQL for sensing devices and StreamSQL for information stream handling. In addition to SQL (Structured Query Language) turned out to be excessively mind boggling because of the persistent augmentations as new abilities are included, designers for the different applications imagined for IoT will think that its difficult to gain proficiency with the majority of SQL’s vernaculars and stunts while they may require just a subset of them. Consequently, it has been recommended that a progressively adaptable structure be utilized, in which a SQL lingo or predefined arrangement is picked by the particular prerequisites of the current situation. Also, a move from particular programming to include situated writing computer programs is proposed to alter programming that is utilized to get to databases. This empowers the improvement of adaptable information the board contingent upon the basic framework’s ideal highlights (Bowman et al. 2007; Sen and Ramamritham 2005). 5.3.3.2  Proficient Handling Procedure The two primary motivations behind gathering information through the “Things” surface in IoT frameworks for announcing and examination. This duo includes handling information sooner or later in the framework so as to total/gather helpful data. Deciding the area at which information is to be prepared in IoT frameworks is an essential structure worry that should think about the decentralized idea of the framework and the volume of information delivered. Two preparing methodologies can be conveyed for IoT frameworks: in-organize handling and incorporated preparing. In-arrange handling includes making out space for the “program” downwards to the information and projecting back the

100

A. Sahani et al.

outcomes to the user(s), accordingly diminishing the magnitude of information which should be moved to concentrated capacity at top building layers in the framework. Incorporated handling, then again, necessitates that information—either in its crude structure or in a totaled, progressively smaller structure—be shipped to persevering stockpiling to empower advanced investigation errands. A half breed of the two methods can be utilized for an increasingly adaptable handling mode, with shifting customizable degrees to oblige assorted needs of applications that run on IoT. 5.3.3.3  Versatile Query Handling, Streamlining and Optimization Query preparing is customarily performed close information stores so as to give inquiry execution plans, which are essentially information getting plans. Conventional question enhancement includes appointing an expense to every one of the various designs for getting information so as to pick the arrangement that is least exorbitant. With regards to IoT, question handling also includes discovering plans to bring information from the sensing devices, routers as well as gadgets that are geologically circulated and returning total readings as results. Along these lines, there is a need for huge chunks of information as messages and trades to convert into correspondence overhead. Bringing down this overhead includes relocating inquiry preparing nearer to things which are paired in lower layer, and receiving temporary improvement to represent restricted questions. Querying improvement for WSNs ought to be seen broadly exploring in writing for WSNs are the most eminent subsystem of Web of Things (IoT); an inside and out and thorough knowledge could be searched for. Improvement in the queries and optimal management which includes picking out the desirable inquiry plans dependent for their individual opinion expenses which gets investigated for WSNs systems. Detecting and directing information and metadata gathering are fused into a query analyzer so as to locate the best inquiry plan with low energy efficient costs. Improvements that are accomplished to the unaccompanied question do not consider the dynamicity of the system. In, fundamental database tasks, for example, read/compose and join activities are allocated power profiles, and an inquiry plan is allocated out a power cost as an element of the activities used to execute that particular plan. Adjusting a comparable methodology for IoT inquiries will include information of the power profiles of middle of the hubs and lower level sensors so as to accomplish the maximum capacity of intensity advanced question (query) plans for questions that are not restricted to information living disconnected of servers. Vitality proficient streamlining of various questions is investigated in ecoDB for appropriated huge scale database frameworks. Bunch enquiry handling uses the nearness of regular segments in different questions, inside a task at hand. This is accomplished by lining inquiries and not executing them right away. Regular sub-­ articulations that might be available in the lined inquiries are then combined and executed. This multi-question improvement accomplishes investment funds in

5  Data Management for the Internet of Things

101

v­ itality utilization to the detriment of an expanded normal reaction time coming about because of the express postponements. Multi-inquiry improvement was likewise utilized for vitality proficient question preparing in Wireless Sensor Medium. The two levels are utilized to advance the course of action of questions: A sink node (main station) level revamps about lot of inquiries to deliver an engineered set of inquiries with redundancies evacuated and regular sub-questions consolidated. Various in-organize improvements are then used to advance the transmission of the inquiry set outcomes. Each of the three strategies has been utilized to accomplish in-organize improvement: time sharing between worldly range questions, having common regular measurement among comparative inquiries through readings communicate, and accumulation of consequence out coming from the sensors reacting for the queries from the sink-node (base station) that has successfully send the manufactured inquiries. Conglomeration Supportability The ceaselessly adaptive nature of information created from tech-savvy items reduces attainability to put resources into mass information gathering and capacity at one store just like the case in traditional DBMS systems. When this is done, information of time-delicate nature will have lapsed. Subsequently, it is attractive to take into consideration for the information to stockpile and occurrence that is continuously progressed as information is created. Be that as it may, this makes the test of broadcasting chunks and huge quantity of information to the facilities where storage is available. Conglomeration as well as combination of sensor information is basic for bringing down the correspondence overhead and anticipated from transmitting crude raw streaming information. Moreover, accumulation and combination might be prerequisites for specific applications where raw information stockpiling has no additional worth. One sound cause that might have thought is about the potential loss of precision coming about because of dropping hidden detailed information. In this way, quick collection ought to be a plan factor for IoT information management systems. Management of Data Framework Systems for IoT The greater part of the present information recommendations are focused on WSNs, that are just a part of the worldwide IoT system, and in this way don’t expressly address the more complex structural attributes of IoT. WSNs are a full grown systems administration worldview whose information arrangements rotate for the most part around in-organize information preparing and enhancement. Sensing devices are generally of constant aligned objects which are asset compelled in nature, which does not encourage modern investigation and administrations. The principle focus in WSN-centric information the executive arrangements are to gather ongoing information immediately for speedy basic leadership, with constrained perpetual capacity capacities with regards to long haul utilization. This speaks to just a subset of the more adaptable IoT framework, which targets saddling the information accessible from an assortment of sources; stationary and versatile, shrewd and installed, asset obliged and asset rich, ongoing and documented. The

102

A. Sahani et al.

fundamental focal point of IoT-based information the board thusly expands the arrangements made for WSNs to include arrangements of a consistent method to take advantage of the volumes of heterogeneous information so as to discover intriguing worldwide examples and key chances (Pujolle 2006; Wu et al. 2011). A mix of information from heterogeneous systems is done in such a way that adjustment and consistent mix of other IoT subsystems is accomplished. Strong and complete information the executive arrangements that help interoperability between assorted subsystems and incorporate the general lifecycle of information the board with the nearness of versatile items and setting mindfulness prerequisites are yet to be created. Here we put forward a system for the Internet of things information the executives which is progressively perfect in the IoT information cycle and approach the structure natives talked about before. The proposed structure includes a overlapping layered methodology which focuses only in the information and information-driven middleware, and gets ideas from the combination of data repository and then board the frameworks to ensure the self-governance of autonomous IoT sub-spaces just as adaptable join/leave engineering. System Representation The IoT information and the executives structure comprises of six(6) stacked layers, out of which two are incorporated layers under the main system and rest are the integrated layers. Here the structure layers maps near the periods of the IoT information lifecycle depicted in, with query/coordination viewed as an additional procedure that isn’t carefully a piece of the information lifecycle. The “Things” Layer incorporates IoT sensors and shrewd items (information generation objects), just as modules for in-arrange preparing and information accumulation/constant collection (handling, total). The Communication Layer offers help for transmission of solicitations, inquiries, information, and results (gathering and conveyance). The information dual layers separately take care of the revelation along with classification of information and the capacity and ordering of gathered (information stockpiling/documented). The Data Layer additionally handles information and inquiry preparing for neighborhood, independent information store locales (separating, preprocessing, preparing). The Federation Layer gives the deliberation and reconciliation of information vaults that is fundamental for worldwide question/investigation demands, utilizing metadata put away in the Data Sources layer to help ongoing mix of sources just as area driven solicitations (preprocessing, coordination, combination). The Querying Layer takes the responsibilities of subtleties for inquiry, preparing as well as streamlining the participation along with the Federation Layer forming a part of the integral Transactions Layer (preparing, conveyance).

5  Data Management for the Internet of Things

103

References Agarwal, P., & Alam, M. (2018). Investigating IoT middleware platforms for smart application development. arXiv preprint arXiv:1810.12292. Bowman, I. T., Bumbulis, P., Farrar, D., Goel, A. K., Lucier, B., Nica, A., Paulley, G. N., Smirnios, J., & Young-Lai, M. (2007, April 17–20). SQL anywhere: A holistic approach to database self-­ management. In Proceedings of IEEE 23rd International Conference on Data Engineering Workshop (ICDE 2007), Istanbul, Turkey, pp. 414–423. Cattell, R. (2010). Scalable SQL and NoSQL data stores. ACM SIGMOD Record, 39, 12–27. Chen, L., Tseng, M., & Lian, X. (2010). Development of foundation models for Internet of Things. Frontiers of Computer Science in China, 4, 376–385. Cooper, J., & James, A. (2009). Challenges for database management in the Internet of Things. IETE Technical Review, 26, 320–329. Guinard, D., Trifa, V., Karnouskos, S., Spiess, P., & Savio, D. (2010). Interacting with the SOA-­ based Internet of Things: Discovery, query, selection, and on-demand provisioning of web services. IEEE Transactions on Services Computing, 3, 223–235. Güting, R. H., & Mamoulis, N. (2011). Special issue on data management for mobile services. VLDB Journal, 20, 641–642. Hauer, J.-H., Handziski, V., Köpke, A., Willig, A., & Wolisz, A. (2008). A component framework for content-based publish/subscribe in sensor networks. In R. Verdone (Ed.), Lecture notes in computer science: Wireless sensor networks (pp. 369–385). Berlin/Heidelberg: Springer. Ozsu, M. T., & Valduriez, P. (2011). Principles of distributed database systems (3rd ed.). New York: Springer. Pujolle, G. (2006, October 3–6). An autonomic-oriented architecture for the Internet of Things. In Proceedings of IEEE John Vincent Atanasoff international symposium on modern computing (JVA 2006), Sofia, Bulgaria, pp. 163–168. Pungila, C., Fortis, T.-F., & Aritoni, O. (2009). Benchmarking database systems for the requirements of sensor readings. IETE Technical Review, 26, 342–349. Ramakrishnan, R., & Gehrke, J.  (2002). Database management systems (3rd ed.). New  York: McGraw-Hill. Sen, R., & Ramamritham, K. (2005, April 5–8). Efficient data management on lightweight computing devices. In Proceedings of the International Conference on Data Engineering (ICDE 2005), Tokyo, Japan, pp. 419–420. Spiess, P., Karnouskos, S., Guinard, D., Savio, D., Baecker, O., Souza, L., & Trifa, V. (2009, July 6–10). SOA-based integration of the Internet of Things in enterprise services. In Proceedings of IEEE International Conference on Web Services (ICWS 2009), Los Angeles, CA, USA, pp. 968–975. Vermesan, O., Harrison, M., Vogt, H., Kalaboukas, K., Tomasella, M., Wouters, K., Gusmeroli, S., & Haller, S. (2009). Internet of Things strategic research roadmap. Brussels: IoT European Research Cluster. Wu, G., Talwar, S., Johnsson, K., Himayat, N., & Johnson, K. D. (2011). M2M: From mobile to embedded internet. IEEE Communications Magazine, 49, 36–43.

Chapter 6

Machine Learning for IoT Systems Ahmed Khattab and Nouran Youssry

Abstract  The rapid increase in the number of smart devices hosting sophisticated applications is significantly affecting the landscape of the information com-­ munication technology industry. The Internet of Things (IoT) is gaining popularity and importance in man’s everyday life. However, the IoT challenges also increase with its evolution. The urge for IoT improvement and continuous enhancement becomes more important. Machine learning techniques are recently being exploit-­ed within IoT systems to leverage their potential. This chapter comprehensively surveys of the use of algorithms that exploit machine learning in IoT systems. We classify such machine learning-based IoT algorithms into those which provide ef-­ficient solutions to the IoT basic operation challenges, such as localization, clus-­tering, routing and data aggregation, and those which target performance-related challenges, such as congestion control, fault detection, resource management and security. Keywords  Internet of Things (IoT) · Wireless sensor network (WSN) · Machine learning · Unsupervised learning · Supervised learning · Fuzzy logic

6.1  Introduction The Internet of Things (IoT) is a networking paradigm that offers pervasive and distributed services in the move towards ubiquitous computing. IoT is a network of objects or things that communicate with each other and with the surrounding environment and share information through the Internet. IoT enables millions of devices, including sensors and smart phones/devices, to be connected for performing different tasks. According to the International Data Cooperation (IDC), the number of

A. Khattab (*) · N. Youssry Electronics and Electrical Communications Engineering Department, Cairo University, Giza, Egypt e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_6

105

106

A. Khattab and N. Youssry

IoT devices worldwide will exceed 50 billion by 2020 producing more than 60 ZB of data (Van der 2017; Sam 2016). Wireless Sensor Networks (WSNs) are playing a main role in IoT. WSNs have attracted significant attention in recent years. A WSN is composed of a set of application-­specific sensor nodes equipped with communication modules. Such nodes gather data from their environment to monitor and record target conditions at diverse locations. While there are sensors that measure almost every environmental aspect, the widely monitored parameters are air temperature and humidity, wind speed and direction, illumination intensity, flow pressure, vibration intensity, sound intensity, power-line voltage, pollution levels, chemical concentrations, and vital body functions. WSNs are powerful in developing application-specific systems. WSNs joins the IoT networking paradigm when the sensor nodes dynamically connect to the Internet to cooperate to achieve their tasks. Both IoT and WSNs face several challenges and issues that should be addressed. Examples include energy efficiency, node localization and clustering, event scheduling, route establishment, data aggregation, fault detection and data security. Exploiting machine learning provides solutions to such problems. Machine learning could significantly boost the performance and distributive characteristic of IoT. Machine learning (ML) emerged as an artificial intelligence (AI) technique in the late 1950s (Ayodele 2010). Since then, its algorithms gradually evolved to become more robust, effective and accurate. Recently, ML classification and regressing techniques have been widely exploited to improve the performance of many of application domains such as bioinformatics, facial and speech recognition, agriculture monitoring, fraud detection and marketing. Machine learning could be used to improve the performance of various IoT systems by exploiting the history of the collected data of given tasks to autonomously optimize the performance without the need to re-program the system. More specifically, the main reasons that make ML important in IoT applications are: • The rapidly changing dynamic nature of the environments typically monitored by IoT systems. Therefore, developing IoT systems that efficiently operate by autonomously adapting to such changes is required. • The unreachable and dangerous settings in which exploratory IoT applications, such as wastewater and volcano eruption monitoring, operate to collect new knowledge. Consequently, the ability of ML-based IoT systems to self-calibrate to the acquired new knowledge is needed to ensure robustness. • Machine learning does not only improve the autonomous control in IoT applications but also ameliorate their intelligent decision-making capabilities. Nevertheless, the use of ML in IoT still face several challenges that should be carefully considered. For instance, IoT devices are resource limited. Using ML to extract consensus relationships between the collected data samples and predicting the accurate hypotheses significantly drain the energy of the IoT devices. This necessitates trading-off the ML algorithm’s computational complexity and the targeted accuracy of the learning process.

6  Machine Learning for IoT Systems

107

In this chapter, we present a brief introduction of IoT: its concept, history, architecture and its processing data layers. We also present a comprehensive survey of machine learning techniques classifying them into five categories which are ­supervised learning, unsupervised learning, reinforcement learning, evolutionary computational and fuzzy logic techniques. We present a detailed study of the applications of machine learning techniques in solving IoT challenges. Moreover, we classify those applications into operational applications which are concerned with the main functions of the IoT system and performance applications which are more concerned with enhancing the performance of IoT systems. The remainder of the chapter is organized as follows. Section 6.2 briefly introduces IoT. Section 6.3 overviews of the different machine learning algorithms. The role of machine learning in solving operational and performance challenges in IoT and WSN systems are discussed in Sects. 6.4 and 6.5, respectively. Finally, our conclusions are drawn in Sect. 6.6.

6.2  IoT Overview IoT is a network of objects “things” which sense, accumulate and transfer data over the Internet without any human intervention. Kevin Ashton, British technology pioneer and co-founder of MIT’s Auto-ID Center, first used the term “Internet of Things” in 1999. Ashton used the term to illustrate the power of using Radio-­ Frequency Identification (RFID) tags to connect goods to the Internet, then count and track them without needing human intervention. Since then, the idea of globally connecting computers and servers through the Internet has been expanded to the Internet of Things in which anything can be connected and accessed through the Internet. This created a whole new connectivity dimension where anything can be connected at anytime and anyplace (Vashi et al. 2017). The industries that are adopting IoT are expected to achieve revenue growth of 22% (Kotha and Gupta 2018). Several IoT architectures have been developed to handle the use of heterogeneous devices in such systems. The number and type of used devices, the application and the amount of collected and processed data control the choice of the most suitable architecture to be used. One simple IoT architectural model is the three-layer model shown in Fig. 6.1 which consists of perception, network and application layers (Ghasempour 2019). • Perception Layer: This is the physical layer of the IoT system. It is composed of the sensors which gather information about the environment and actuators which implement the actions that accordingly change the environment. A temperature controller in an air conditioner is an example of an actuator. • Network layer: This is the transmission layer that is responsible for handling the routing decisions. It is also handling the transmission and processing of the data received from or transmitted to the perception layer.

108

A. Khattab and N. Youssry

Fig. 6.1  IoT architecture models

• Application Layer: This layer delivers application-specific services to end users. It also provides the interface between humans and the IoT system. Such three-layer architecture represents the simplest IoT architecture. As the data size of the system increase, this architecture becomes inefficient. That is why the five-layer model shown in Fig. 6.1 was proposed in (Sethi and Sarangi 2017), adding the following three layers to the perception and application layers: • Transport Layer: This layer handovers data from the perception layer to the processing layer and actions in the reverse direction. Several network types are used for this purpose such as wireless LAN, NB-IoT, LoRA, RFID, and NFC. • Processing Layer: This is the middleware layer that stores and processes the huge amounts of data received from the transport layer. It also prepares the data for the application layer. The processing layer manages and provides a wide range of services to the lower layers. Cloud computing, databases and big data analysis are examples of the technologies used in this layer. • Business Layer: This layer encompasses the overall IoT application alongside its business and profit models. It is also responsible for the end users’ privacy and security. Another architecture presented in (Navani et al. 2017) proposed the same division of layers with only changing their names. The layers in this architecture are the object, object abstraction, service management, application and business layers. Cloud computing was originally used to implement the processing layer because it provides significant flexibility and scalability. A cloud database management system based on a five-layer architecture was introduced in (Alam et  al. 2013). Significant efforts were carried out to enhance cloud computing system’s database with query processing mechanism in (Malhotra et al. 2018) and to enhance the task scheduler as in (Ali et al. 2019). As energy is known to be a scarce resource in IoT

6  Machine Learning for IoT Systems

109

systems, the authors of (Ali and Alam 2016) proposed energy management techniques for cloud computing environments. IoT devices generate valuable data readings that need to be transferred. Therefore, merging cloud computing technology with big data analysis (Alam and Shakil 2016) is a very important in many platforms as discussed in (Khan et al. 2016, 2018, 2019a, b, c, d; Shakil et al. 2017). Lately, the increase of real-time applications requiring the least possible latency caused a migration towards another processing architectures that involve either fog or edge computing. In fog computing paradigms, the data and its processing take place in decentralized computing structures that physically reside between the data sources and the cloud. Hence, fog computing results in low-latency and is suitable for time-sensitive IoT applications as data is processed close to where it is originated. Its efficiency is also higher as less data is uploaded to the cloud. On the other hand, data processing takes place either on the IoT device generating the data or on a local gateway device that resides in the vicinity of the IoT device in the edge computing paradigm. Thus, both fog and edge computing reduce the dependence on the cloud infrastructure in data analysis, which in turn reduce the system latency, and hence, allow the data-driven decisions making process much faster. However, fog computing is the better option where data aggregation from different sources is needed whereas edge computing is better where the least latency is allowed. The differences between cloud computing, fog computing and edge computing are illustrated in Fig. 6.2.

Fig. 6.2  Mapping IoT processing layers to system devices

110

A. Khattab and N. Youssry

6.3  Machine Learning Taxonomy Machine learning techniques are designed to automatically benefit from prior experience in acting in the future without explicit reprogramming. Existing ML approaches are typically classified as either supervised, unsupervised or reinforcement learning. However, artificial intelligence techniques have recently played a great role in enhancing ML techniques. Therefore, this chapter categorizes ML techniques into supervised learning, unsupervised learning, reinforced learning, evolutionary computation and fuzzy logic. This section briefly overviews the dif-­ ferent ML approaches in addition to their most updated algorithms in the context of IoT and WSNs.

6.3.1  Supervised Learning In supervised learning, the input and targeted output data are both labeled for classification. This presents the learning base on which future data processing is centered. The key supervised learning algorithms are: 1. k-nearest Neighbor (k-NN): In this supervised learning approach, a data sample is classified according to the labels of nearby data samples. Simple methods (e.g., the Euclidean distance between the IoT devices) are typically used to compute the average measurements of neighboring devices within a specific range. It is a simple computational algorithm but may be inaccurate in large data sets. In IoT, k-NN is used in fault detection (Warriach and Tei 2017) and data aggregation approaches (Li and Parker 2014). 2. Support Vector Machine (SVM): Decision planes are used in SVM approaches to define decision boundaries. A decision plane separates groups of objects each with different class memberships as depicted by the example shown in Fig. 6.3. Fig. 6.3  Support vector machine example

6  Machine Learning for IoT Systems

111

SVM supervised learning is typically used for localization problems (Kang et al. 2018) to detect malicious behaviors and to address several security issues in IoT and WSNs (Zidi et al. 2018) due to its high accuracy. 3. Neural Network (NN): An artificial neural network (ANN) imitates biological neurons by interconnecting layers of artificial neurons. These artificial neurons map the different sets of input data onto a set of appropriate outputs. Figure 6.4 presents the model of a neural network. Even though ANNs provide solutions to non-linear and complex problems, they are computationally complex. Artificial NNs improve the efficiency of IoT localization (Banihashemian et al. 2018; El Assaf et al. 2016), detect faulty nodes (Chanak and Banerjee 2016), and establish routing (Mehmood et al. 2017). 4. Bayesian Interface: Unlike most machine learning algorithms, Bayesian inference uses a reasonably small number of samples for training. Bayesian methods efficiently learn uncertain perceptions by adapting probability distributions while avoiding overfitting. However, they need prior knowledge about the environment. Bayesian interface is suitable for fault detection (Warriach and Tei 2017), cluster head selection (Jafarizadeh et al. 2017) and localization (Sun et al. 2017; Wang et al. 2017; Guo et al. 2018) approaches. 5. Decision Tree (DT): A tree-like model of decision or classification is used in such a decision-support tool. DTs are created using a set of if-then conditions. For boosting DT accuracy, the random forest (RF) algorithm is introduced. RF is an ensemble decision tree method that operates by constructing multiple classifiers. Each classifier is a decision tree. RFs are used in intrusion detection as in (Varsha et al. 2017).

6.3.2  Unsupervised Learning Unsupervised learning algorithms operate over datasets in which the input data does not have labeled responses. Inferences are drawn in such algorithms by classifying the unlabeled input data into groups that are called clusters. 1. Principal Component Analysis (PCA): PCA reduces the dimension of a large set of variables to a much smaller set that contains almost all the information in the Fig. 6.4  Neural networks’ structure

112

A. Khattab and N. Youssry

original large set. PCA application in IoT systems for data dimensionality reduction takes place either at the sensor or the cluster head levels. PCA results in a reduction in the communication overhead (Wang et al. 2019) which is very useful in data aggregation (Liu et al. 2017). 2. k-means Clustering: It is used to classify different data in classes or clusters (Jain et al. 2018). k random centroids are initially chosen. The other nodes then join the clusters of the nearest centroid. Averaging over the nodes in each cluster, new centroids are determined. The algorithm repeats the previous steps until convergence is reached. 3. Self-Organizing Maps (SOM): A self-organizing map is also considered as a method for dimensionality reduction as explained in (Miljković 2017). However, a SOM is a type of artificial neural network which result in a discretized low-­ dimension representation of the input data, called a map, using unsupervised learning for training. SOMs are very suitable to be used in building clusters in IoT.

6.3.3  Reinforcement Learning Reinforcement learning (RL) does not have knowledge about the inputs nor their corresponding outputs. It is a very important ML technique whose idea, illustrated in Fig. 6.5, is that an agent will learn from the environment by interacting with it and receiving rewards for performing actions. Over the past few years, RL algorithms have been used for designing routing protocols in IoT systems and WSNs to reduce the energy consumption and improve the network performance (Habib et al. 2018). An extensively used RL algorithm is Q-learning. First, a Q table is initialized, then an action a is performed. A reward is then measured to update the Q table. In order to assess how good to take a certain action a at a particular state s, the action-­ value function Q(s, a) is learnt by the algorithm. Initially, the action a is randomly chosen until the Q table is constructed, then the best action is chosen from it.

Fig. 6.5 Reinforcement learning concept

6  Machine Learning for IoT Systems

113

6.3.4  Evolutionary Computation Unlike other ML approaches, evolutionary computation techniques solve problems using computational models that mimic the biological behavior of either humans or animals in problem solving tasks. 1. Genetic Algorithms (GA): Such algorithms use biologically inspired heuristic search techniques to find the best solution for problems with large search spaces. GAs work in parallel on a population of solutions rather than processing a single solution. First, a chromosome structure is defined, typically in the form of an array of bits as shown in Fig. 6.6. Then, an initial chromosomes population is randomly generated for which fitness is evaluated. Chromosomes with higher fit-ness are selected in the selection process. A crossover process combines two parents to introduce a new child to the population. Finally, mutation randomly up-dates the parents to introduce new children. An example GA cycle is shown in Fig. 6.6. GAs are suitable for data aggregation approaches and searching for optimal clusters. 2. Ant Colony Optimization (ACO): ACO probabilistically searches for the optimal path in a graph in ways similar to how ants find the path between a food source and the colony. First, ants move randomly, leaving traces or pheromone on the taken path. More pheromone on a path indicates that the path probability to be the shortest/optimum one is high. This algorithm is efficient for routing in IoT. 3. Particle Swarm Optimization (PSO): PSO is inspired by swarm theory, fish schooling, and bird flocking. As an evolutionary computation approach, PSO searches for the best solution in a population. The algorithm starts with a random

Fig. 6.6  Genetic algorithm example

114

A. Khattab and N. Youssry

population of solutions referred to as particles. A fitness function is used to compute the particle’s fitness value which is optimized in each generation. IoT clustering algorithms recently exploit PSO to improve their performance.

6.3.5  Fuzzy Logic Traditional ML techniques used to work with binary values: either 0 (False) and 1 (True). However, fuzzy logic (FL) imitates the way of decision making in a human which considers all the possibilities between 0 and 1 digital values (Umarikar 2003). FL introduces the concept of degree of truth. Its value does not have to be exactly 1. It can be any real value between 0 and 1 instead. Fuzzy logic is an attractive solution for localization typically used to combine the node’s residual energy, centrality, and distance from the data sink node for electing the best cluster heads (Umarikar 2003).

6.4  Machine Learning for IoT Basic Operation In this chapter, we categorize the challenges that face IoT systems into basic operation and performance-related challenges. In this section, we take a closer look on how ML is making an effective contribution in solving the basic system operation challenges such as node localization, clusters formation, routing, and data aggregation. Figure 6.7 summarizes how ML is used in addressing such challenges.

6.4.1  Node Localization The procedure of determining the geographic coordinates of the nodes is known as localization. As much as it is crucial to be aware of IoT nodes’ locations, it is impractical to use GPS hardware in every node as it dramatically consumes the nodes’ energy. Alternatively, localization can exploit machine learning alongside some parameters such as the received signal strength (RSS) and the time and angle of arrival. An approach was introduced in (Sun et al. 2018) to use neural networks in localizing WSN nodes. The proposed solution uses the variations of the RSS between the sensor nodes. RSS is measured from all the nodes once without the presence of any target, then with the target presence. The ANN uses the difference in RSS values and the corresponding matrix indices as inputs. The ANN outputs are the nodes’ locations. The ANN is trained to approximate a nonlinear function to map the inputs and outputs. The use of SVM in localization was proposed in (Kim et al. 2013). What makes this approach different is the use of ensemble SVM technique. The ensemble ­technique employs multiple ML techniques, then decides the result by voting. In

6  Machine Learning for IoT Systems

115

Fig. 6.7  Machine learning exploitation for IoT basic operation

(Kim et al. 2013), the authors used multiple SVMs by dividing the WSN to many subnetworks. Each subnetwork is trained using SVM. The resulting sub-predictions are then combined as ensemble combination. Each training problem has a smaller size compared to the size of the original problem. Consequently, better results are obtained. Another advantage of having fewer sensor nodes per subnetwork is the reduction in the transmission power and communication energy as the nodes become close to each other. The idea of treating localization problem as regression problem in-stead of a classification one was introduced in (Bhatti 2018). The algorithm is divided into two phases. The training dataset for localization in WSN contains of the feature vectors related to each anchor and its true location coordinates which are already known. The feature vector of an anchor node is composed of the RSS values of the signals received from other nodes as measured by that anchor node. In the second phase, the sensor nodes’ coordinates are estimated. The input of the learned model is the target WSN nodes’ feature vectors. The produced output is the nodes’ coordinates. Comparing the extended feature vector (readings from anchor and ­sensor nodes) versus the reduced feature vectors (readings from anchor vectors only) showed that the extended features provide a better prediction accuracy.

116

A. Khattab and N. Youssry

Another localization technique was explained in (Kumar 2018) using fuzzy logic ML technique. The proposed technique was a hybrid Sugeno-Mamdani fuzzy system using RSS for localization which outperformed the traditional fuzzy logic. The authors then proposed the idea of cooperative localization. In such scheme, whenever an unknown node gets localized, it acts as anchor or landmark for the next iteration and transmits beacons to other unknown nodes. Since it has already been localized now and can broadcast beacons which contain its position, identity to be received by other unknown nodes. A low-complexity centroid-based scheme is proposed in (Kumar 2018). The resulting precision error is high. Hence, the authors used a FL algorithm to improve the nodes’ location estimation based on FL weights. For more optimization, the consequences of unbalanced node placements in non-­ uniform networks are alleviated using PSO. This approach proved its efficiency in networks with few nodes and limited sensing coverage. As the number of nodes and/or the sensing coverage increase, integrating extreme machine learning tech-­ niques (as NN) with the centroid scheme showed better results in (Phoemphon et al. 2018).

6.4.2  Clustering IoT systems are energy-constrained networks. Transmitting all the data packets to the sink node is inefficient and dramatically consumes the nodes’ energy. A local aggregator, or a cluster head (CH), is used to improve the energy-efficiency by collecting the data from the cluster members within its vicinity and transmitting only the aggregation of the data to the sink node. Machine learning algorithms can help in deciding the number of clusters needed and electing the cluster heads. An integrated approach in which clustering is performed using a SOM phase followed by a k-means phase was introduced in (Ahmadnezhad and Rezaee 2015). The SOM input parameters are the energy levels and the nodes’ coordinates. The weight vectors of the SOM map units are selected nodes with maximum energy levels. Such maximum-energy nodes attract the nearest lower-energy nodes, thereby, creating energy-balanced clusters. Fuzzy logic techniques can also be used in electing cluster heads as proved in (Nayak and Devulapalli 2016). The authors of (Nayak and Devulapalli 2016) proposed a fuzzy logic clustering approach using the battery power, node mobility and node centrality as input parameters to a FL system that finds the probability of a node to serve as a cluster head. Simulation results showed that the fuzzy logic cluster head election system outperforms the well-known LEACH clustering protocol in terms of the network lifetime defined as the time until first node dies, last node dies or half the nodes die. A cluster formulation method in which a node individually decides its ability to serve as a CH rather than executing an election process is presented in (Forster and Murphy 2006). This clustering method exploits Q-learning alongside a set of dynamic parameters such as the nodes’ energy levels. A reinforcement learning technique was also used in (Soni and Shrivastava 2018) to implement an on-demand

6  Machine Learning for IoT Systems

117

mobile sink traversal. Recently, evolutionary computation algorithms are used showing enhancement in solving the clustering problem as in (Wang et al. 2018) where ACO-based approach is used. Likewise, Energy Centers Searching using Particle Swarm Optimization (EC-PSO) is proposed in (Wang et al. 2019). A geometric method is initially used to elect the CHs. Then, EC-PSO performs clustering when the nodes’ energies start to be heterogeneous. EC-PSO elects the nodes close to the energy center to be CHs using an improved PSO technique that searches the energy centers. PSO algorithm increases the network lifetime as proved in (Yadav et al. 2018).

6.4.3  Routing Designing a routing protocol for IoT systems is very challenging due to their nature of restricted processing, compact memory, and low bandwidth. Routing protocols address several issues including scalability, energy utilization, data coverage, and fault tolerance while optimizing their tradeoffs. Machine learning can effectively address this challenge as it continuously discovers the optimal routing paths that result in the best tradeoffs is the dynamically changing IoT networks. ML also reduces the complexity of a typical routing problem by breaking it down to subrouting problems that only consider the local neighbors of the nodes. Finally, ML effectively achieves the routing QoS requirements despite the use of computationally inexpensive algorithms and classifiers (Alsheikh et al. 2014). A wireless routing protocol that uses SOM alongside a modified radial-based neural network was presented in (Hoomod and Jebur 2018). It starts with clustering the networks using SOM as previously explained. Then, an ANN will be responsible for finding the optimal path. However, the used ANN is modified by having the weights to the output layer computed and attuned using the parallel Moore-Penrose generalized pseudo-inverse which accelerates the learning process and accuracy. Taking time as a comparison metric, the proposed protocol outperformed the traditional Dijkstra in fixed and mobile topologies. FL was also used in solving the routing challenge. In (Mottaghinia and Ghaffari 2018), two fuzzy logic-based systems were proposed to route data messages and specify their priority. When a source node encounters other nodes, it checks the data delivery probability of each node alongside the node’s energy. A node is not considered as a potential router if either the residual energy and delivery probability is low, or both is low. Ultimately, the fuzzy system will select the best candidate for data transmission from the nodes’ neighbors. The proposed approach shows its efficiency by increasing the data delivery rate and decreasing the associated delay. Another approach combines FL (for clustering) and ACO (for routing) (Arjunan and Sujatha 2018). The node’s residual energy, degree, centrality and distance to both the Base Station (BS) and neighboring nodes are used as inputs to the FL-based clustering algorithm which maximizes the network lifetime while balancing the nodes’ energies. ACO is used to get the shortest paths. Also, periodic random choice

118

A. Khattab and N. Youssry

of paths occurs to exploit unused paths and to balance the nodes’ energies. Again, FL is used in (Thangaramya et al. 2019) but this time for enhancing the NN to be used in discovering energy-efficient routes. The NN discovers new routes by investigating the consumed energy in the nodes and the routing patterns. The routes’ weights are attuned by applying FL rules to reach the most efficient route.

6.4.4  Data Aggregation Data aggregation is very crucial to reduce the IoT system power consumption. It combines and summarizes the data packets of several nodes properly at the cluster head. Data aggregation decreases the number of transmitted data packets, thereby, increases the bandwidth utilization and minimizes the energy consumption. In what follows, we demonstrate the power of machine learning techniques in this field. First, (Pinto et al. 2014) proposed using genetic algorithms in information fusion to perform a trade-off between the Quality of Fusion (QoF) and efficiency by dynamically adjusting the sending probability. According to the effect of a node’s input on the system performance, each node is given a reward that will decide whether to send data or fuse it with another node instead. A priority-based data aggregation approach was introduced in (Ghate and Vijayakumar 2018). It works in supervised mode if the input data has class labels. For example, the health issue or the disease may be known previously in certain cases, hence, it can be used as a class label in the output vector. Alternatively, this approach may work in unsupervised mode with input data lacking class labels. A novel approach combining PCA and Angle Optimized Global Embedding (AOGE) was introduced to tackle the concept drift problem (Liu et al. 2017). In ML context, concept drift implies that the target variable, to be predicted by the model, has time-­ varying statistical properties that vary in unforeseen ways. Consequently, the prediction process results in less accurate predictions with time. AOGE takes advantage of several techniques. The projection variance of sampled data is first analyzed. Then, PCA is used to define the dispersion in the data. The principal components are then selected considering the maximized projection variance. Unlike PCA, AOGE analyzes the projection angle of sampled data to choose the principal components. Consequently, AOGE outperforms PCA when tested using real-life datasets with significantly noisy data. This implies that even though PCA and AOGE separately detect concept drift in a data stream, their combination is more effective and robust in detecting concept drift.

6  Machine Learning for IoT Systems

119

6.5  Machine Learning for IoT Performance Aspects While the basic operational challenges are directly associated with functional behavior of IoT systems, performance aspects are mostly associated with performance enhancement. The performance enhancing requirements include fault detection, mitigation and controlling congestion provide quality of service and maintain security. This section sheds the light on the exploitation of machine learning in such performance-related aspects (summarized in Fig. 6.8).

6.5.1  Congestion Control Congestion negatively impacts the performance of IoT applications as it causes packet losses, increases the encountered delays, wastes the nodes’ energies and significantly degrades the IoT application fidelity. The purpose of IoT and WSN congestion control is to improve the network throughput and reduce the time of data

Fig. 6.8  Machine learning exploitation for enhancing IoT performance

120

A. Khattab and N. Youssry

transmission delay. Thus motivated, the authors of (Alaa 2018) proposed to have a congestion detection phase followed by a congestion monitoring phase. The system detects the congestion through measuring the data loss rate. The congestion monitoring system is simply an ANN for learning about the congestion scenarios to reduce and stop them before they even occur. This approach proved significant improvement when compared against the traditional case of not having congestion control.

6.5.2  Fault Detection As explained in (Warriach and Tei 2017), faults occur when one or more of the IoT system characteristics or parameters deviate from their normal operation or value. Faults occur when node is faulty because of a physical damage, a low battery, communication interference, or environmental interference. An error is defined as an incorrect sensing of a state or event in the given space due to a fault. Faults can be classified into: –– Offset fault: This fault occurs when the data always differ from its expected value by a constant amount because of faulty calibration of the sensing module. –– Gain fault: This fault occurs when the rate of change of the sensed data in a period of time differs from its expected value. –– Stuck-at fault: This type of faults happens when the sensed data is constant and does not vary with time (zero-variance). –– Out-of-bounds fault: This fault happens when the values of the sensed data exceeds the normal operation bounds. In (Warriach and Tei 2017), the fault detection problem is changed into a simple classification problem where the received data either belongs to a normal or a defected class. Three machine learning approaches were used for this purpose: k-­ NN, SVM and Naïve Bayes. k-NN has the least classification error in the least computing time followed by SVM. However, Naïve Bayes showed the worst performance. In (Zidi et al. 2018), the authors brought attention to a different kind of faults and how to solve it. This fault was the random fault that is defined as an instant error in which data is disturbed just for an instant of time. This error causes many positive or negative sharp peaks that influence the data of the sensors. These perturbations are very fast which makes them more difficult to detect. The authors proposed an SVM classifier for detecting instant errors which showed a high accuracy that reached 99%. Evolution of the traditional classifiers is proposed in (Javaid et  al. 2019) by implementing Enhanced SVM (ESVM) which combines SVM and GA.  Also, the authors implemented Enhanced KNN (EKNN) and Enhanced Recurrent Extreme Learning Machine (ERELM) which gives the most accurate results. Another classifier was introduced in (Noshad et al. 2019) which uses RF algorithm and shows better results compared to SVM and NN.

6  Machine Learning for IoT Systems

121

6.5.3  Resource Management To satisfy the enormous resource demands of the various IoT applications, robust resource management techniques are needed to minimize the energy consumption and the response time. Since IoT systems are dynamic in nature, RL is one of the most suitable ML technique in IoT resource management as proposed in (Kumar and Krishna 2018). However, RL complexity increases with the increase in action pairs. Researchers combined NN and RL to introduce Drift Adaptive Deep Reinforcement Learning (DA-DRL) in (Chowdhury et al. 2019) to enhance traditional RL methods. Another scheduling technique was suggested in (Zhang et al. 2019) as Q-Learning Scheduling on Time Division Multiple Access (QS-TDMA) to improve the real-time reliability.

6.5.4  Security As IoT systems are resource limited, securing such systems against security at-tacks presents an immense challenge. Several approaches for secure authentication in IoT systems through cloud computing exist (Kumari et  al. 2018; Alam et  al. 2015). However, the majority of contemporary IoT commercial devices suffer severe security flaws and vulnerabilities as shown in (Williams et al. 2017). That is why the demand for using ML techniques is rapidly growing to save such networks from different security attacks. Here, we discuss the major five IoT attacks (Mamdouh et al. 2018). –– Distributed Denial of Service (DDOS) Attack: In this cyber-attack, the attackers overload the system making it difficult to be used by its intended users by sending multiple requests to exceed its capacity, and therefore, crushing. –– Spoofing Attack: Is a cyber-attack where attackers aim to masquerade and deceive the system by pretending to be an authorized node to trick them in performing legitimate actions or giving up sensitive data. –– Malware Attack: Is a cyber-attack in which a malware or a malicious software performs activities on the victim’s operating system, usually without the node’s knowledge. –– User to Root (U2R): In this attack, the attacker attempts to escalate a user’s privilege from being limited to become a super user or be able to access the root. This is achieved using stolen credentials or through a malware infection. –– Remote to Local (R2L): In this attack, the attacker imitates a legitimate user to gain remote access to a victim device. In (Doshi et al. 2018), the authors detect DDOS attacks through capturing the data traffic and analyzing its features. It was proved that normal IoT traffic differs from DDOS traffic in terms of packet size, inter packet arrival, used protocol, the bandwidth and node/IP destination. Based on that observation, normal ML

122

A. Khattab and N. Youssry

c­ lassification techniques such as SVM, k-NN, ANN and Random Forest were used. It was shown that Random Forest and k-NN perform best. However, an updated technique was presented in (Thamilarasu and Chawla 2019) which combines ML and deep learning algorithms to form a deep neural network to detect DDOS attacks more precisely. NN is used with cooperation on cloud trace back technique to detect DDOS attacks as in (Alam et al. 2015). Detecting spoofing attacks needs four main stages as illustrated in Fig. 6.9. ML techniques are commonly used in the feature detection and attack detection stages. In (de Lima Pinto et al. 2018), feature detection was achieved using a k-means algorithm and a k-NN classifier was proposed. Another technique was proposed in (Pajouh et al. 2019) that uses PCA in dimension reduction in addition to two classifiers: Naïve Bayes classifier followed by a k-NN classifier. This two-tier classification has high detection rates and accurate detection of the U2R and R2L hard-to-detect security attacks. Malware detection was approached as a classification problem using random forest and k-NN in (Pajouh et  al. 2019). An interesting approach for intrusion detection was shown in (Yu and Tsai 2008) in which each sensor node is equipped with an intrusion detection agent (IDA). As nodes cannot trust each other, IDAs do not cooperate. A Local Intrusion Detection Component (LIDC) is responsible for extracting the local features such as the packet delivery and collision rates, delays, number of neighbors, cost of routing and consumed energy. Meanwhile, Packet based Intrusion Detection Component (PIDC) infers if a suspected node is launching an attack on the host by analyzing the suspected node’s packets and investigate the packets’ RSS, the arrival rate of the sensed data and retransmission rate of the attacker’s packets. Then, SLIPPER machine learning algorithm is used for detection. Finally, securing WSN and IoT middleware using Generative Adversarial Networks (GANs) was proposed in (Alshinina and Elleithy 2018). The proposed approach is composed of two networks. A generator network that generates fake data that mimics the real sensed data and confuses the attacker by combining both the fake and real data. A discriminator network is then used to separate the fake data from the real data. This does not only protect data from adversaries but also improves the data accuracy compared to conventional techniques.

Fig. 6.9  The detection stages of spoofing attacks

6  Machine Learning for IoT Systems

123

6.6  Concluding Remarks The unique nature of WSNs and IoT systems gives us no choice but to address their challenges and limitations through suitable tools and specified techniques. Here comes the need for machine learning techniques either supervised learning, unsupervised learning, reinforcement learning, evolutionary computation or fuzzy logic. All such techniques offer different solutions to most of the challenges. In this chapter, we have discussed these solutions for addressing the IoT basic operation challenges such as node localization, cluster formulation, routing and data aggregation. Moreover, we have discussed the role of machine learning in solving the performance-­related challenges such as congestion control, fault detection, resource management and security. We conclude with the following remarks: • Performance related aspects mainly exploit supervised learning techniques. Such challenges are handled as classification tasks where the algorithm needs to predict discrete value or identify the input data into a particular class. This requires prior knowledge and that is why supervised ML techniques are suitable here. • Evolutionary techniques are used for solving basic operation rather than performance related aspects. Their objective is injecting new actions and measuring their effect e.g. by imitating ants in reaching their destination. This makes evolutionary technique not suitable for performance challenges which are typically modelled as classification tasks. • Since fuzzy systems are capable of handling uncertainties and giving wide range of truth, they are recently being adopted for IoT routing and node localization. • Resource management is solved using reinforcement learning (Q-learning technique). IoT systems are very dynamic. Managing their resources also needs a dynamic technique that always interacting with the surrounding environment to make the right immediate actions. Hence, RL comes as a perfect match here.

References Ahmadnezhad, F., & Rezaee, A. (2015). Increasing the lifetime of wireless sensor networks by self-­ organizing map algorithm. International Journal of Computer Networks and Communications Security, 3(4), 156–163. Alaa, M. (2018). Radial basis neural network controller to solve congestion in wireless sensor networks. Iraqi Journal for Computers and Informatics, 44(1), 53–62. Alam, M., & Shakil, K.  A. (2016). Big data analytics in cloud environment using Hadoop. In International conferences on mathematics, physics & allied sciences. Alam, B., Doja, M., Alam, M., & Malhotra, S. (2013). 5-layered architecture of cloud database management system. AASRI Procedia Journal, 5, 194–199. Alam, M., Shakil, K.A., Javed, M.  S., & Ansari, M. (2015). Ambreen: Detect and filter traffic attack through cloud trace back and neural network. In International conference of parallel and distributed computing.

124

A. Khattab and N. Youssry

Ali, S. A., & Alam, M. (2016). A relative study of task scheduling algorithms in cloud computing environment. In 2nd international conference on contemporary computing and informatics (IC3I). Ali, S. A., Affan, M., & Alam, M. (2019). A study of efficient energy management techniques for cloud computing environment. In 9th international conference on cloud computing, data science & engineering (Confluence). Alsheikh, M., Lin, S., Niyato, D., & Tan, H. (2014). Machine learning in wireless sensor networks: Algorithms, strategies, and applications. IEEE Communication Surveys and Tutorials, 16(4), 1996–2018. Alshinina, R., & Elleithy, K. (2018). A highly accurate deep learning based approach for developing wireless sensor network middleware. IEEE Access, 6, 29885–29898. Arjunan, S., & Sujatha, P. (2018). Lifetime maximization of wireless sensor network using fuzzy based unequal clustering and ACO based routing hybrid protocol. Applied Intelligence, 48(8), 2229–2246. Ayodele, T. (2010, February). Introduction to machine learning. In Y. Zhang (Ed.), New advances in machine learning. IntechOpen. Banihashemian, S., Adibnia, F., & Sarram, M. (2018). A new range-free and storage-efficient localization algorithm using neural networks in wireless sensor networks. Wireless Personal Communications, 98(1), 1547–1568. Bhatti, G. (2018). Machine learning based localization in large-scale wireless sensor networks. Sensors, 18(12), E4179. Chanak, P., & Banerjee, I. (2016). Fuzzy rule-based faulty node classification and management scheme for large scale wireless sensor networks. Expert Systems with Applications, 45(C), 307–321. Chowdhury, A., Raut, S., & Narman, H. (2019). DA-DRLS: Drift adaptive deep reinforcement learning based scheduling for IoT resource management. Journal of Network and Computer Applications, 138, 51–65. de Lima Pinto, E., Lachowski, R., Pellenz, M., Penna, M., & Souza, R. (2018). A machine learning approach for detecting spoofing attacks in wireless sensor networks. In IEEE international conference on Advanced Information Networking and Applications (AINA). Doshi, R., Apthorpe, N., & Feamster, N. (2018). Machine learning DDoS detection for consumer Internet of Things devices. arXiv preprint arXiv: 1804.04159. El Assaf, A., Zaidi, S., Affes, S., & Kandil, N. (2016). Robust ANNs-based WSN localization in the presence of anisotropic signal attenuation. IEEE Wireless Communications Letters, 5(5), 504–507. Forster, A., & Murphy, A. (2006). CLIQUE: Role-free clustering with Q-learning for wireless sensor networks. In IEEE international conference on distributed computing systems. Ghasempour, A. (2019). Internet of Things in smart grid: Architecture, applications, services, key technologies, and challenges. Inventions, 4(1), 22. Ghate, V., & Vijayakumar, V. (2018). Machine learning for data aggregation in WSN: A survey. International Journal of Pure and Applied Mathematics, 118(24), 1–12. Guo, Y., Sun, B., Li, N., & Fang, D. (2018). Variational bayesian inference-based counting and localization for off-grid targets with faulty prior information in wireless sensor networks. IEEE Transactions on Communications, 66(3), 1273–1283. Habib, A., Arafat, M., & Moh, S. (2018). Routing protocols based on reinforcement learning for wireless sensor networks: A comparative study. Journal of Advanced Research in Dynamical and Control Systems, (14), 427–435. http://www.jardcs.org/backissues/abstract. php?archiveid=6166 Hoomod, H., & Jebur, T. (2018). Applying self-organizing map and modified radial based neural network for clustering and routing optimal path in wireless network. Journal of Physics: Conference Series, 1003, 012040. Jafarizadeh, V., Keshavarzi, A., & Derikvand, T. (2017). Efficient cluster head selection using naïve bayes classifier for wireless sensor networks. Wireless Networks, 23(3), 779–785.

6  Machine Learning for IoT Systems

125

Jain, B., Brar, G., & Malhotra, J. (2018). EKMT-k-means clustering algorithmic solution for low energy consumption for wireless sensor networks based on minimum mean distance from base station. In Networking communication and data knowledge engineering. Springer. Javaid, A., Javaid, A., Wadud, Z., Saba, T., Sheta, O., Saleem, M., & Alzahrani, M. (2019). Machine learning algorithms and fault detection for improved belief function based decision fusion in wireless sensor networks. Sensors, 19(6), 1334. Kang, J., Park, Y., Lee, J., Wang, S., & Eom, D. (2018). Novel leakage detection by ensemble CNNSVM and graph-based localization in water distribution systems. IEEE Transactions on Industrial Electronics, 65(5), 4279–4289. Khan, S., Shakil, K. A., & Alam, M. (2016). Educational intelligence: Applying cloud-based big data analytics to the Indian education sector. In 2nd international conference on contemporary computing and informatics (IC3I). Khan, S., Shakil, K. A., Ali, S. A., & Alam, M. (2018). On designing a generic framework for big data-as-a-service. In: IEEE international conference on advanced research in engineering sciences. Khan, S., Shakil, K. A., & Alam, M. (2019a). PABED – A tool for big education data analysis. In 20th IEEE international conference on industrial technology. Khan, S., Liu, X., Ara Shakil, K., & Alam, M. (2019b). Big data technology – Enabled analytical solution for quality assessment of higher education systems. International Journal of Advanced Computer Science and Applications (IJACSA), 10(6). ESCI/Scopus. Khan, S., Arshad Ali, S., Hasan, N., Ara Shakil, K., & Alam, M. (2019c). Big data scientific workflows in the cloud: Challenges and future prospects. Cloud Computing for Geospatial Big Data Analytics, 1–28. Khan, S., Shakil, K. A., Alam M. (2019d). Big data computing using cloud-based technologies: Challenges and future perspectives. Networks of the Future: Architectures, Technologies and Implementations. Kim, W., Park, J., Yoo, J., Kim, H., & Park, C. (2013). Target localization using ensemble support vector regression in wireless sensor networks. IEEE Transactions on Cybernetics, 43(4), 1189–1198. Kotha, H., & Gupta, V. (2018). IoT application – A survey. International Journal of Engineering & Technology, 7, 891–896. Kumar, A. (2018). A hybrid fuzzy system based cooperative scalable and secured localization scheme for wireless sensor networks.. International Journal of Wireless & Mobile Networks (Vol. 10, pp. 51–68). Kumar, T., & Krishna, P. (2018). Power modelling of sensors for IoT using reinforcement learning. International Journal of Advanced Intelligence Paradigms, 10(1–2), 3. Kumari, A., Abbasi, M. Y., Kumar, V., & Alam, M. (2018). The cryptanalysis of a secure authentication scheme based on elliptic curve cryptography for IOT and cloud servers. In IEEE International Conference on Advances in Computing, Communication Control and Networking (ICACCCN). Li, Y., & Parker, L. (2014). Nearest neighbor imputation using spatial–temporal correlations in wireless sensor networks. Information Fusion, 15, 64–79. Liu, S., Feng, L., Wu, J., Hou, G., & Han, G. (2017). Concept drift detection for data stream learning based on angle optimized global embedding and principal component analysis in sensor networks. Computers and Electrical Engineering, 58, 327–336. Malhotra, S., Doja, M. N., Alam, B., & Alam, M. (2018). Generalized query processing mechanism in cloud database management system. In Big data analytics (pp. 641–648). Singapore: Springer. Mamdouh, M., Elrukhsi, M., & Khattab, A. (2018). Securing the Internet of Things and wireless sensor networks via machine learning: A survey. In IEEE International Conference on Computer and Applications (ICCA). Mehmood, A., Lv, Z., Lloret, J., & Umar, M. (2017). ELDC: An artificial neural network based energy-efficient and robust routing scheme for pollution monitoring in WSNs. IEEE

126

A. Khattab and N. Youssry

Transactions on Emerging Topics in Computing, 1–1. https://ieeexplore.ieee.org/abstract/ document/7859382/citations#citations Miljković, D. (2017). Brief review of self-organizing maps. In IEEE international convention on information and communication technology, electronics and microelectronics (MIPRO). Mottaghinia, Z., & Ghaffari, A. (2018). Fuzzy logic based distance and energy-aware routing protocol in delay-tolerant mobile sensor networks. 100(3): 957–976. Navani, D., Jain, S., & Nehra, M. (2017). The Internet of Things (IoT): A study of architectural elements. In 13th international conference on Signal-Image Technology & Internet-Based Systems (SITIS). Nayak, P., & Devulapalli, A. (2016). A fuzzy logic-based clustering algorithm for WSN to extend the network lifetime. IEEE Sensors Journal, 16(1), 137–144. Noshad, Z., Javaid, N., Saba, T., Wadud, Z., Saleem, M., Alzahrani, M., & Sheta, O. (2019). Fault detection in wireless sensor networks through the random forest classifier. Sensors, 19(7), 1568. Pajouh, H., Javidan, R., Khayami, R., Dehghantanha, A., & Choo, R. (2019). A two-layer dimension reduction and two-tier classification model for anomaly-based intrusion detection in IoT backbone networks. IEEE Transactions on Emerging Topics in Computing, 7(2), 314–323. Phoemphon, S., So-In, C., & Niyato, D. (2018). A hybrid model using fuzzy logic and an extreme learning machine with vector particle swarm optimization for wireless sensor network localization. Applied Soft Computing, 65, 101–120. Pinto, A., Montez, C., Araújo, G., Vasques, F., & Portugal, P. (2014). An approach to implement data fusion techniques in wireless sensor networks using genetic machine learning algorithms. Information Fusion, 17, 90–101. Sam, S. (2016). Internet of Things’ connected devices to triple by 2021, reaching over 46 billion units. Juniper Research. Sethi, P., & Sarangi, S. (2017). Internet of Things: Architectures, protocols, and applications. Journal of Electrical and Computer Engineering, 1, 1–25. Shakil, K. A., Zareen, F. J., Alam, M., & Jabin, S. (2017). BAM health cloud: A biometric authentication and data management system for healthcare data in cloud. Journal of King Saud University  – Computer and Information Sciences (in press). https://www.sciencedirect.com/ science/article/pii/S1319157817301143 Soni, S., & Shrivastava, M. (2018). Novel learning algorithms for efficient mobile sink data collection using reinforcement learning in wireless sensor network. Wireless Communications and Mobile Computing, 2018:7560167, 13 pages. Sun, B., Guo, Y., Li, N., & Fang, D. (2017). Multiple target counting and localization using variational Bayesian EM algorithm in wireless sensor networks. IEEE Transactions on Communications, 65(7), 2985–2998. Sun, Y., Zhang, X., & Wang, X. (2018). Device-free wireless localization using artificial neural networks in wireless sensor networks. Wireless Communications and Mobile Computing, 2018, 4201367, 8 pages. Thamilarasu, G., & Chawla, S. (2019). Towards deep-learning-driven intrusion detection for the internet of things. Sensors, 19(9), 1977. Thangaramya, K., Kulothungan, K., Logambigai, R., Selvi, M., Ganapathy, S., & Kannan, A. (2019). Energy aware cluster and neuro-fuzzy based routing algorithm for wireless sensor networks in IoT. Computer Networks, 151, 211–223. Umarikar, A. (2003). Fuzzy logic and brief overview of its applications. University Västerås Suecia. Van der Meulen, R. (2017). Gartner says 8.4 billion connected things will be in use in 2017, up 31 percent from 2016. Garther Research. Varsha, S., Shubha, P., & Avanish, T. (2017). Intrusion detection using data mining with correlation. In 2nd international conference for Convergence in Technology (I2CT). Vashi, S., Ram, J., Modi, J., Verma, S., & Prakash C. (2017). Internet of Things (IoT): A vision, architectural elements, and security issues. In IEEE international conference on IoT in Social, Mobile, Analytics and Cloud (I-SMAC).

6  Machine Learning for IoT Systems

127

Wang, Z., Liu, H., Xu, S., Bu, X., & An, J. (2017). Bayesian device-free localization and tracking in a binary RF sensor network. Sensors, 17(5), 1–21. Wang, J., Cao, J., Sherratt, R., & Park, J. (2018). An improved ant colony optimization-based approach with mobile sink for wireless sensor networks. The Journal of Supercomputing, 74, 6633–6645. Wang, J., Gao, Y., Liu, W., Sangaiah, A., & Kim, H. (2019). An improved routing schema with special clustering using PSO algorithm for heterogeneous wireless sensor network. Sensors, 19(3), 671. Warriach, E., & Tei, K. (2017). A comparative analysis of machine learning algorithms for faults detection in wireless sensor networks. International Journal of Sensor Networks, 24(1), 1–13. Williams, R., McMahon, E., Samtani, S., Patton, M., & Chen, H. (2017). Identifying vulnerabilities of consumer Internet of Things (IoT) devices: A scalable approach. In IEEE international conference on Intelligence and Security Informatics (ISI). Yadav, A., Kumar, S., & Vijendra, S. (2018). Network life time analysis of WSNs using particle swarm optimization. Elsevier, 132, 805–815. Yu, Z., & Tsai, J. (2008). A framework of machine learning based intrusion detection for wireless sensor networks. In IEEE international conference on sensor networks, ubiquitous, and trustworthy computing. Zhang, B., Wu, W., Bi1, X., & Wang, Y. (2019). A task scheduling algorithm based on Q-learning for WSNs. The Abel Prize, 521–530. Zidi, S., Moulahi, T., & Alaya, B. (2018). Fault detection in wireless sensor networks through SVM classifier. IEEE Sensors Journal, 18(1), 340–347.

Chapter 7

Supervising Data Transmission Services Using Secure Cloud Based Validation and Admittance Control Mechanism Kamta Nath Mishra

Abstract  In the recent era cloud computing is being extensively used for numerous services. Since, adoption has security concerns. Therefore, this research work presents attestation innards to evaluate the metamorphosis utility of mobile and electronic health data transmission through cloud network and it better comprehends the idea of infiltration trialing and susceptibility assessment. The proposed approach of this research work contains information on common threats to cloud based mobile health services and their corresponding solutions. The idea of this research work is to utilize Internet Communication Technology (ICT), Internet of Things (IoT) and Embedded Computing Technology (ECT) to enhance the performance of secure cloud (SecC) based m-health data transmission services. In this research work the author has explored a new and secure cloud based service which can upgrade the concert of cloud based m-health transmission systems. Here, an advanced internet communication machinery (ICM) based on IP-WSN (Internet Protocol-Wireless Sensor Network), scattered scheme architectures, web-computing, wisdom and motivation is used for enhancing the performance of cloud based m-health data communication systems. The cloud server of proposed m-health system will also provide access and control mechanism for multiple types of vehicles on roads and hence the proposed system will be helpful in enhancing the safety of passengers while they are travelling from their houses to the e-health hospitals. The services and safety of cloud based m-health data transmission services can be optimized in real-time situations. Keywords  Cloud computing · e-health · Embedded computing technology · Internet communication technology · m-health data transmission services · Secure cloud

K. N. Mishra (*) Department of Computer Science & Engineering, Birla Institute of Technology, Ranchi, Jharkhand, India © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_7

129

130

K. N. Mishra

7.1  Introduction The web Computing is a web dependent latest computing system which is allocating facilities in a favor of clients e.g. platform computing infrastructure, sharing network resources, and software products. The Cloud benefit actualize in this dynamic market fragment are: Amazon, Microsoft, Google, IBM, Oracle, Eucalyptus, VMware, Eucalyptus, Citrix, Sales power and rack space just as there are a wide range of trader offering distinctive cloud administrations (Albeshri and Caelli 2010). With the enormous increase of cloud and web crime based services, the terms of suitable web system defense system has become essential in order to look after integrity, confidentiality, and accessibility of transportation facilities. The necessities of real time transmission in the pulling strategy and bring into play of ongoing systems administration aptitude for all customers winds up imperative to keep up adaptability and Quality of Service (QoS) in web systems. If we need to provide user friendly accessibility, reliability and secrecy of essential data positioned on a web computing network, a variety of safekeeping measures will be needed. If the cloud wants to retain a stable and secure web network, the web server requires all the essential measures to be implemented, the faithful access privileges have to be implemented and outer boundaries are to be appropriately defined (Sedigh et al. 2014; Alger 2005). Many researchers, industry experts and academicians have tried to define the term web computing and its characteristics. The researchers Buyya et al. (2008) and Khodadadi et al. (2015) have defined it as the cloud is a collection of interconnected virtualized dynamically provisioned computers which provide parallel and distributed computing environment and it presents one or more unified computing resources established through negotiation between the service provider and consumers. Van Bon et al. (2007) described that clouds are a large pool of easily usable/ accessible virtualized resources which are dynamically reconfigured to adjust to a variable load (scale) and it optimizes resource utilization. The pool of resources in cloud computing environment is shared according to pay-per-use model. Here, guarantees are offered by the infrastructure provider as per the service level agreements. Miller (2008) claimed that clouds are hardware based services where computation, network and storage capacity are offered with highly elastic infrastructure capacity. A report from the University of California says that the key characteristics of cloud computing are infinite computing resources, up-front commitment elimination and pay per use. Similarly, the researcher Alger D. (2005) said that cloud is more often used to refer as the information technology infrastructure deployed on an IaaS data center. Here, a resource means compute application, system assets, platform, software services, virtual servers and compute infrastructure. Web Computing is being widely used in many industries. The Cloud computing has a lot of usage in many areas like accounting applications, customer relationship management, communications and collaboration, electronic mails and shared calendars,

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

131

financial management, office productivity suits, online storage management, human resource and employment etc. The cloud computing has various advantage like 24/7 support, reimburse as much as you use, scalability, high level computing, virtualized and self-motivated atmosphere (Mell and Grance 2009; Khalid et al. 2013). The services and safety of cloud based transmission system can be optimized in real-time situations. The SeC with ICT can take an interest a vital role to set up link among different servers of the clouds and end users to improve accident preclusion. When in doubt, the assurance is a joint trustworthiness of the web customer and web server as the records well being and secrecy issue trick genuine trepidation (Plummer et al. 2009; Singh et al. 2015a). To stay away from these critical issues, the creator’s present a legitimate and idiot proof confirmation strategy to inspect the execution of pulling administrations (Singh et  al. 2015b). The explanation in web compute merges amenity safety, amenity supervision, and observes of resources. In the current era no standard rules and regulations are existing for deploying applications in cloud based systems. Further, there is a serious be deficient in consistency power in the web environment. Various new processes have been planned and imposed in web. But, these presented techniques are not enough to ensure total security because of the dynamics of web environment. The intrinsic issues of data safety, management, supremacy and control of cloud data are discuss in (Mahmood 2011). Sun et al. (2011) described the key protection, isolation, and dependence issues in the accessible environment of cloud computing and provided help to the users to identify the substantial and insubstantial fear related to cloud computing. As per the author’s opinion, there are three core possible terrorization namely safety, isolation and faith in cloud computing, here, safety plays a decisive role in the existing era of cloud and distributed computing. The cloud security can further be categorized into four subcategories namely safety mechanism, data confidentiality, web server monitoring, and preventing illegal operations & service hijacks. A records safety structure for web computing networks was anticipated by Pandey et al. (2013). Here, the researchers mostly discuss the protection concerns linked to the storage of cloud data. Various patent concerning the security of statistics techniques for cloud and distributed systems were filed by (Klein 2013). Younis and Kifayat provided a analysis on web security supervision for critical infrastructures (Younis and Kifayat 2013). A security and isolation structure for radio frequency identification (RFID) in web computing environment was projected by Kardaş et al. (2013) which efficiently integrates RFID technology with internet of things (IoT) using web computing environment. In a nutshell, it very well may be said that the main issues of web data insurance comprise of records security, records openness, data assurance, and safe data correspondence. The fundamental security challenges in distributed computing condition incorporate danger location, recuperation of information misfortune, benefit disturbance aversion, discovery and avoidance of pernicious assaults (Behl 2011). The scientists Chen and Zhao (2012) investigated information security and protection issues in the cloud and conveyed processing condition by concentrating on

132

K. N. Mishra

protection insurance, cloud security and information isolation. Subsequently, it tends to be said that the information security issues are principally at SaaS, PaaS, and IaaS levels. Further, the most extreme test in distributed computing is i­ nformation sharing. The association of information security and protection issues in distributed computing condition is exhibited in Fig. 7.1. In this research work the author has explored a new and secure cloud based service which can upgrade the concert of cloud based transmission systems. Here, advanced internet communication machinery (ICM) based on IP-WSN (Internet Protocol-Wireless Sensor Network), dispersed framework structures, distributed computing, detecting and activating is utilized for upgrading the execution of cloud based information correspondence frameworks. The cloud server of proposed system will also provide access and control mechanism for multiple types of vehicles on roads and hence the proposed system will be helpful in enhancing the safety of passengers. The remaining parts of this research work are organized as follows. The Sect. 7.2 describes the objectives and challenges of cloud based research. The area 7.3 shows a short discourse about distributed computing condition and its security worries for transportation framework. This area likewise depicts a safe cloud based transportation administrations. The segment 7.4 presents secure cloud test-bed setup model and shows of proposed design are assessed in this segment. At long last, the creator has finished up the substance of this exploration work in area 7.5.

Public cloud Cloud

Hybrid cloud

Private cloud

Data security and privacy

Data integrity Data confidentiality Data availability

Software

Hardware

Data privacy

Fig. 7.1  Association of information security and protection issues in distributed computing condition

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

133

7.2  O  bjectives and Challenges of Cloud Based Data Transmission Research However, virtualization and web computing provide a wide variety of active resources. The problem of protection is professed specifically as the big problem on the web, which causes users to oppose utilizing Distributed computing advancement. Security is the rule worry of the cloud; cloud master focuses and clients facing huge security issues.

7.2.1  Objectives The fundamental target of this examination is to comprehend security dangers and recognize the suitable security systems used to moderate them in Cloud Computing administrations, to comprehend cloud security issues and strategies utilized in the present Cloud world. Distinguishing security in the difficulties of the cloud, it is normal that later on of distributed computing it will recommend measures to counter the future difficulties that distributed computing administrations confront. As we probably are aware, in an open cloud framework we have a foundation underlined with numerous virtual machines. A few associations utilize virtual machines to get to their secret information.

7.2.2  Challenges The cloud computing environment addresses the challenges of next generation cloud computing architectures. It also addresses the challenges of permitting application platforms and development platforms to take advantage & reimbursement of web computing. The examined on web computing is tranquil in the beginning period. Many presented issue have yet not been address whereas the new challenge are emerging on daily basis. Several of these challenges are: Service Level Agreements (SLA’s), Web Data Management & Security, Data Encoding, and Relocation of virtual Machines and Server Consolidation (Wijaya 2011; Handley and Rescorla 2006). Further, this world needs to find the answer of following questions which are related to cloud based m-health data transmission and security (Hogan et al. 2011; Almulla and Chon 2010; Tan et al. 2011): • What are the different security systems which are being utilized by the main cloud specialist co-ops when the information is being transmitted among the mists and neighborhood m-health systems?

134

K. N. Mishra

• What is the variety of safety technique being used to stop unlawful entrance of m-health data within the web? • What are the key safety issue and challenge which the world is expecting in future about m-health data transmission in cloud computing environment? • How the world can tackle these safety troubles which are predictable to arise in potential web processor systems for m-health data transmission?

7.3  A  ccess Control Mechanisms of Data Transmission Using Cloud Services The access Control is a control access list is prepared in-order to specify and control who can access & what can be accessed. As Cloud computing is used by multiple users and multiple enterprises and heterogeneous infrastructure, so new cloud safety concerns arrive which are Multi-tenancy, Velocity of Attack, Information Assurance and Data Privacy and Ownership. To make the cloud infrastructure secure, the users need to use various cloud safety mechanism. In a virtual environment, the computer, network and storage are virtualized so, to maintain safety. The Fig. 7.2 shows cloud safety method which is applicable for all cloud based m-health communication systems (http://searchsecurity.techtarget.com/definition/authentication-authorization-and-accounting; Ahmed et al. 2013; Hogan et al. 2011).

Fig. 7.2  The cloud safety method

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

135

7.3.1  Securing Virtual Machines The virtual machines (VMs) can be secured by hardening the virtual machines. There can be one more component which can be the point of attack is Hypervisor Management System. This software will manage the hypervisor. Securing virtual machines (VMs) may further include virtual machine (VM) isolation and VM hardening. If one virtual machine has been attacked and compromised in the cloud infrastructure then that VM need to be isolated from the rest of the VM’s to make sure that this attacker does not get control over the rest of the virtual machine or the infrastructure. Hardening is the method of shifting the evasion arrangement in organizes to accomplish superior safety. By limiting the resource that VM can consume to prevent denial of services attacks, disable unused functions and devices on virtual machine, use a directory service for authentication; perform vulnerability scanning and penetration testing of the guest operating system secures virtual machine (Albeshri and Caelli 2010; Ahuja 2011; Bellare et al. 2001, 2002). The operating system needs to be updated very regularly and the safety software’s are up to date. The applications also need to be updated regularly so that they are secure. The identity management is one more concern as public cloud infrastructure has multiple user belonging to multiple organizations and ensuring authorization that the user is who the user claims to be so by adding multiple layers of safety like multifactor authentication and biometric authentication, can make sure that only authorize users can login to our system (Alger 2005; http://cloudcomputing.sys-con.com/node/612375/print).

7.4  G  overnance and Operations in Cloud Computing Environment The domination refers to the policy, process, law and institution that describe the configuration by which company are intended for and manage. Each organization will have its safety policies which will have a different set of policies at its information technology level so making sure that these policies are enforced correctly on a cloud infrastructure is also important. There are also external regulations like banking sector has its own set of laws making sure that the confidential data of the user is not compromised, healthcare industry has its own set of laws. All these government regulations have to be followed by the company else the company can be prosecuted. Internal and external regulations are also followed. In case of a traditional data centre, the IT department of that company will take care of it. In case of cloud, the cloud service provider along with the organization’s IT department has to ensure that the organizations policies are also applied in cloud environment (Alger 2005). Now, in these days web computing process can challenge several conformity examination necessities at present in place. Statistics location; web computing

136

K. N. Mishra

Fig. 7.3  The overview of governance and operations on m-heath in cloud networks

safety guidelines lucidity; and IAM, are all tough issues in agreement audit efforts. The example of the agreement requirements include confidentiality and law; Payment Card Industry (PCI) necessities; and monetary coverage laws. It is a fear of company about the resiliency of web computing, since statistics may be commingled and scattered roughly different servers and biological region (Alani 2010; Whiting et al. 2002). It may be achievable that the statistics for a precise certain point cannot be recognized. Distinct recent swarm, the venture know accurately where the position is of their statistics, to be fast retrieve in the event of failure healing. In the web computing representations, the principal Cloud examination supplier might farm out capability to third parties, who might too farm out the healing process. This will become more difficult when the primary Cloud service provider does not ultimately hold the data. The Figs. 7.3 and 7.4 show the overview of cloud safety and cloud security (SeC) framework of cloud based network systems (Whiting et  al. 2002; Chadwick and Fatema 2012; Saldhana et al. 2014).

7.5  S  ecurity Issues for Cloud Servers and Transmission Systems Web processing is generally utilized in different ventures and is encountering noteworthy development in around humankind. The important peculiarity of web processing is to a great extent access to the framework; in requires self-benefit, quick

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

137

Fig. 7.4  The SecC framework for m-health data transmission in cloud networks

adaptability, reservation booking and exhaustive examination. Passage to the immense framework is the contact of a few properties facilitated in money organizes from a wide scope of areas and offers online access. On-request self-benefit is the availability of administrations and assets from cloud specialist organizations when required. Speedy flexibility gives adaptable administrations dependent on the necessities of the provider. The gathering of assets serves clients and clients through versatile and impermanent administrations. The deliberate administration directs the administrations given by the providers to the purchaser, which incorporate the receipt and adequate utilization of assets. The key three building squares which are usually utilized administration models in distributed computing condition are: Software as A Service (SaaS), Platform as A Service (PaaS), and Infrastructure as A Service (IaaS). Here, SaaS will give the product and applications according to the requests over the web. In this way, the expense of programming, upkeep and activities will be limited. The PaaS will give the stage on interest all through the web for creating programming items utilizing instruments/libraries from the specialist organization. The PaaS will likewise bolster programming organization and arrangement settings. The IaaS will give the framework on interest all through web for data/information preparing, stockpiling and system assets. The association of arrangement model and basic qualities with crucial building squares of distributed computing condition is exhibited in Fig. 7.5 (Whiting et al. 2002; NIST Publication 2001). As cloud is increasing greater fame, an ever increasing number of associations need to move towards cloud yet the key worry about moving towards cloud has been security. Today security is required in every one of the organization models. As per NIST, the cloud display is made out of four noteworthy organization models to be specific Public, Private, Community and Hybrid mists. The general population cloud foundation is free for open use by the overall population of web. The private cloud framework is saved for the advantaged utilization of an association c­ omprising

138

K. N. Mishra

Fig. 7.5  Cloud computing building blocks internaction with other components for m-health data and videos/images transmission

of numerous customers. The people group foundation in the cloud is gotten for an advantaged use by a particular network of buyers who have shared interests, for instance. Mission, governmental issues, and so forth. The mixture cloud foundation is an association of at least two diverse cloud frameworks that have a novel presence, however are connected by an institutionalized innovation that gives the convenience of information and applications (Caballero et al. 2007; Black and Rogaway 2000; Rogaway and Black 2002).

7.5.1  S  ecurity Concerns in Data/Video Transmission Through Cloud Server Most cloud users are unaware of the risks involved in storing and transmitting private information in a shared environment. Therefore, the main technological constraints such as transparency, multiple ownership, attack speed, information security, privacy and data ownership, compliance, cryptography and integrity must be addressed carefully. Therefore, customers are not completely safe or immune from Internet threats and this requires an appropriate secure cloud mechanism and periodic reviews to manage current technologies. To reach the highest level of secure cloud server, it is important that every client (controller) in the cloud network is secure and aware. IETF defines the denial of service attack (DoS) (Singh et  al. 2015b) in which one or more machines attack a victim and try to prevent the victim from doing useful work. In general, DoS attacks can be detected by analyzing the characteristics of the victim’s network. The most effective defenses against DoS and DDoS attacks must filter routers, disable IP transmissions, apply security patches and disable unused services that perform intrusion detection tasks (Robert et  al. 2011; Johnson 1999; Masson and Loftus 2003). Figure 7.6 presents four key key processes for each customer to identify and test individually in the cloud server, and these key processes are: Induction, Research,

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

139

Fig. 7.6  Interaction mechanism of 4PP for transferring client’s/patient’s data in cloud computing environment

Interaction, and Intervention. Induction is the process of understanding the objective by analyzing the environment in which it is located. It is because the reliability of the objectives in their surrounding environment. The induction process is represented by the variable “Z” in Fig. 7.6. The search concept marked as an identifier (“C”) can simply be explained as the search process of the objective foundation. In this process, the objective is analyzed to detect any emanation or any identifier of these emanations. Identified interaction identifier (‘A’ and ‘B’), these interactions are simply answers to questions or agitations initiated by the analyst. From examples of such interactions, ICMP responds and tracks course operations. Finally, the intervention is the process of the analyst who imitates some resources and services that the objective requires for the operation. This process helps to identify the extreme levels in which the target could still operate and is shown by the variables “X”, “Y” and “Z” in Fig. 7.6. Multiple ownership is an important security problem for cloud networks. Here, surface attacks have increased due to the co-location of several virtual machines for a single server.

7.5.2  Transportation System Requirements Today the economy of each country is growing at a tremendous rate. People in all parts of the world have cars to travel. Some buy it as a status symbol, while others buy it as a necessity. We often observe the growing overflow of cars on every road

140

K. N. Mishra

Fig. 7.7  Current transportation system

in developed countries. On the other hand, the car accidents are increasing every day and many people of this earth are forced to leave this earth permanently within fractions of seconds after the occurrence of accident. The Fig.  7.7 shows a road accident which has occurred in the current transportation system. To limit auto collisions, a confirmation instrument for the vehicle framework is introduced in the cloud, displayed in Fig. 7.8. In the vehicle scope of the proposition it is said that the utilization of the Wi-Fi/Bluetooth module associated by means of their individual interfaces to the coordinated processor can transmit pictures and information to the equipped specialists inside the predetermined time and, in this manner, the lives of numerous individuals caught in episodes can be spared. The proposed consistent cloud topology (SeC) of Fig.  7.8 will be introduced inside every vehicle. In this way, the association among vehicles and other traffic control offices will be moved forward. This enhanced and enhanced correspondence and control framework dependent on distributed computing will be extremely valuable to avoid vehicle mishaps on streets and motorways.

7.5.3  Secure Cloud Test Setup Model The setup model of the SeC test to test the cloud-based correspondence framework is displayed in Fig. 7.8. Here, the customer server consistent topology is executed in the test seat. The test setup display organize is intended to interface and coordinate with basic corporate norms. Utilize a various leveled approach that utilizes excess, versatility, and key usefulness dependent on connection total. In the test setup model of Fig.  7.8 there are two associations between the Core and every appropriation switch. All in all, virtual neighborhoods virtual local area networks (VLANs) are utilized in the WAN of the proposed server-based correspondence display in the

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

141

Fig. 7.8  The SecC based logical topology for taking patients to e-health hospitals

cloud. The execution of VLANs in the proposed correspondence show altogether expands time productivity and helps the system head in charge of arranging remote gadgets through Telnet sessions (Shakil et al. 2015; Khan et al. 2017a). To encourage the way toward arranging the VLAN and to bring together this movement, the virtual transport convention (VTP) was utilized halfway as a server and as a customer in whatever is left of the gadgets in the system. In this way, the VLANs are designed in the focal switch and the VTP server running in the part sends intermittent updates over the system to guarantee that every gadget has bare essential information on the setup of the VLAN.  These updates contain an amendment number, so when they land at the customer, these modification numbers can be analyzed and the customer will refresh the data of its VLAN database dependent on that procedure. The two associations between the system center and every conveyance switch are designed as connections to have VLAN traffic. Here, every association underpins the local VLAN to permit the vehicle of label less VLAN traffic over the system, between various sections (Miller 2008; Singh et al. 2015b; Caballero et al. 2007). The creator has set up two passageways to test the organization on two free dispersion level switches. System address interpretation (NAT) has been arranged to secure residential areas to guarantee that it can’t be gotten to through unprotected

142

K. N. Mishra

transport conventions, for example, ICMP. NAT has been designed to enable certain delivers to be unmistakable to the outside world (e.g. Web and record server). All together for the switch to perform steering between VLANs, the EIGRP (Enhanced Internal Gateway Routing Protocol) is favored in this inquiry report. The connection and trade of data between various clients/vehicles of the proposed cloud-based correspondence framework is exhibited in Fig. 7.9. The bearers are the principle correspondence joins for clients with the specialist co-op’s cloud. These vectors have been appeared in Fig. 7.9. Essentially, every one of these vectors can be associated with a different test situation. This enables better outcomes to be accomplished because of its compartmentalized structure, with the goal that the presence of an excessive number of changes can be kept away from. The past advances were taken to finish the data gathering and the test arranging stage. Nonetheless, the checking procedure was not performed independently for every bearer. This examining and list process produces general data on noticeable advancements and transport administrations (Shakil et al. 2017; Khan et al. 2017b). The cloud vector correspondence arrangement of Figs. 7.8 and 7.9 utilize computerized signals for a wide range of correspondence. Regardless, on the off chance that one of the clients utilizes the simple flag, the relating simple signs are changed over

Fig. 7.9  Communication vectors necessity for cloud server based health service providers and patients interaction

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

143

into computerized signals utilizing a simple advanced converter before transmitting the signs (Ali and Alam 2016; Malhotra et al. 2017).

7.5.4  Performance Evaluation The Metasploit is a valuable utilization and weakness testing tool which helps using dividing the saturation testing workflow into smaller and further manageable jobs. With the help of Metasploit the cloud computing researchers can control the power of the Metasploit framework. The Metasploit helps using automating the process of in depth utilization of cloud networks, and provides the required tools to perform manual testing of penetration test cases. The researchers can use Metasploit to scan the open ports and services of cloud computing networks; The Metasploit utilizes vulnerabilities, gathers evidences, and generates the detailed reports of test results. The Metasploit Pro is a multi-user collaborative tool which helps us in the distribution of tasks and information with other group members of a diffusion testing group. In order to perform penetration testing using the team members can share the host’s data, examine the composed evidences, and generate host notes to distribute awareness about any or particular target. The Metasploit was utilized as the fundamental data get-together and helplessness appraisal apparatus to all the more likely comprehend the hidden reasons for vulnerabilities in the executed proving ground. This segment shows the outcomes gathered from the Graphical User Interface (GUI) of Metasploit structure. The outcomes assembled from Metasploit are as reports, yet these reports have excess data which won’t all be utilized for assessment and investigation in this examination work. In any case, just the essential data, pertinent to this idea has been introduced in this examination work. The proposed system was tested in the self created test set-up where clients were allowed to use different types of operating systems for cloud communication. The outputs given by proposed system corresponding to clients, servers, and used operating systems are presented in Fig. 7.10 (Singh et al. 2015b; Caballero et al. 2007; Black and Rogaway 2000). The CDP-Snarf is a cloud computing network sniffer tool which was exclusively written for digging out information from CDP packets. It is a Linux based bundle which is utilized for picking up data about the gadgets in the system and sniffing Cisco Discovery Packets (CDPs). The CDP-Snarf provides all types of information through a command “show cdp neighbors detail”. This command can be executed on any Cisco router. The CDP-Snarf can be easily tested with internet protocol version 4 (IPV4) and IPV6. The information displayed for a cloud computing network using CDP-Snarf are: Time intervals between CDP advertisements, Source MAC address, CDP Version, Checksum, Device ID, Software version, Platform, Addresses, Port ID, Save packets in PCAP dump file format, Read packets from PCAP dump files, and Debugging information. The Fig. 7.11 is showing a screen-­ shot where the proposed cloud based communication system is communicating with network gateway for providing interaction between different clients.

144

K. N. Mishra

Fig. 7.10  OS fingerprints enumeration in the created test set-up

Fig. 7.11  The output obtained from CDPSnarf

The CGE.pl is a bundle of purl that works in a Linux situation. This bundle can possibly over-burden the ports on system gadgets and consequently thinks of them as unusable. Figure 7.12 demonstrates the achievement of the specialized adventures that have been performed in the cloud-based correspondence framework proposed utilizing the CGE.pl device (Singh et al. 2015b; Ahmed et al. 2013; Caballero et al. 2007). Zenmap is a ground-breaking examining apparatus utilized for checking and identifying systems. The Fern WiFi Cracker is a modified apparatus that

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

145

Fig. 7.12  Peak load testing of proposed cloud based e-health communication system using CGE. pl packet

Fig. 7.13  Fern-WiFi-Cracker attack successfully handled in an e-health communication system

utilizes extra bundles to break passwords from remote passageways utilizing distinctive strategies, for example, lexicon documents. In Fig.  7.13 the creator ­demonstrated the catch of the trial leave screen in which the outside assaults on the

146

K. N. Mishra

AP of a Wi-Fi based cloud correspondence framework were effectively overseen by the proposed framework. According to the reports generated, the v2.0 population chains of the Simple Network Management Protocol (SNMP) are vulnerable to exploitation. Therefore, it can be easily broken if used in a cloud communication system. This issue isn’t explicit to a piece of the system however to the whole system. In this manner, approaching the network channels utilized by the SNMP programming likewise implies approaching all and every one of the gadgets associated with the system, including the principle segment of the organization support. Subsequently, the capturing of the SNMP v2 session can have troublesome outcomes. Any accomplished robber can control arrange setups by getting to SNMP certifications and can cause DoS assaults. Thusly, data on the organization system can be acquired illicitly (Singh et al. 2015b; Black and Rogaway 2000; Rogaway and Black 2002; Robert et al. 2011). Another imperative viewpoint to consider would be the development of CDPs into the system. Any abnormal state master/ruffian can undoubtedly decide the open ports and the different functionalities of the gadgets utilizing the data incorporated into these bundles. Programming forms of Cisco gadgets are likewise promoted on the CDP. In this manner, aggressors and thieves can search for adventures that are impressions of the system interconnect working framework (IOS). To counter this damage, the creator of this exploration recommends crippling CDP reports on all Cisco gadgets on the off chance that they don’t take care of the issues. To keep up availability amid such assaults, the creator suggests utilizing directing conventions, for example, the fringe portal informing convention (BGMP) on the grounds that it can keep up numerous doors. Subsequently, availability will be kept up in the cloud-­ based correspondence framework regardless of whether one of the doors is blocked. The Fern-Wifi-Cracker is an unsafe device, since it underpins multi-abuse systems and utilizations modernized word reference documents to split the passwords on remote APs. The Zenmap is an extremely solid data gathering instrument which can be utilized to assemble a gigantic amount of data about the objective system. The significant sweep utilizing Zenmap can distinguish open ports, working framework fingerprints and a partial topology chart. In any case, the Zenmap isn’t a period talented apparatus for system testing. At last, the most grounded instrument for cutting a system is the Metasploit structure which produces reports for further assessment and investigation notwithstanding perform assaults. Accordingly, the similarity test can be allowed to enhance the comprehension of system status (Johnson 1999; Masson and Loftus 2003).

7.6  Conclusions This exploration work plainly demonstrates that the development of innovation produces the requirement for cutting edge safety efforts that are affected by expanded vulnerabilities and complex assaults. On the positive side, the development of inno-

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

147

vation has a negative effect from an aggressor’s perspective. As per the aggressor’s perspective, the absence of top to bottom learning of the framework can be misused in the following procedure. As a rule, the principal finish of this examination work is that arrange security should basically be finished with deceivability. Its significance is that if an individual can conceal their system assets from aggressor gatecrashers, the likelihood of system dangers will be incredibly decreased. This is because of the structure through which infiltration tests are performed. List is one of the underlying phases of every infiltration test. On the off chance that it must be isolated from the cycle, the accompanying advances will be out of date. There is dependably a harmony between security dangers and relating countermeasures, yet these measures by and large focus on recognizing and blocking dangers as opposed to concealing vital system assets from them. Accordingly, it very well may be said that cloud specialist organizations have less straightforwardness than others. Therefore, it can make clashes inside the corporate data the executives’ framework.

References Ahmed, I., James, A., & Singh, D. (2013). Critical analysis of counter mode with cipher block chain message authentication mode protocol—CCMP. Security and Communication Networks, 7(2), 293–308. Ahuja, R. (2011). SLA based scheduler for cloud storage and computational services. International Conference on Computational Science and Applications (ICCSA), pp. 258–262. Alani, M.  M. (2010). Testing randomness of block-ciphers using diehard test. International Journal of Computer Science and Network Security, 10(4), 53–57. Albeshri, A., & Caelli, W. (2010). Mutual protection in a cloud computing environment. 12th IEEE international conference on High performance Computing and Communications (HPCC), pp. 641–646. Alger, D. (2005, June). Build the best data center facility for your business. Indianapolis: Cisco Press Book. Ali, S. A., & Alam, M., (2016). A relative study of task scheduling algorithms in cloud computing environment. In Proceedings of the 2nd international conference on contemporary computing and informatics (IC3I 2016), pp. 105–111. Almulla, S., & Chon, Y.-Y. (2010). Cloud computing security management. 2nd international conference on engineering systems management and its applications, pp. 1–7. Behl, A. (2011). Emerging security challenges in cloud computing: An insight to cloud security challenges and their mitigation. In Proceedings of the World Congress on Information and Communication Technologies (WICT’11), pp. 1–5. Bellare, M., Kilian, J., & Rogaway, P. (2001). The security of the cipher block chaining message authentication code. Journal of Computer and System Sciences, 61(3), 362–399. Bellare, M., Kohno, T., & Namprempre, C. (2002). Authenticated encryption in SSH: Provably fixing the SSH binary packet protocol. In V. Altari, S. Jajodia, & R. Sandhu (Eds.), Proceedings of 9th annual conference on Computer and Communications Security – CCS 2002, held 18–22 November 2002 in Washington, USA (pp. 1–11). New York: ACM Publication. Black, J., & Rogaway, P. (2000). A suggestion for handling arbitrary-length messages with the CBC-MAC.  In M.  Bellare (Ed.), Proceedings of 20th annual international conference of advances in cryptology – CRYPTO 2000, Lecture notes in computer science 1880, held 20–24 August 2000 in Santa Barbara, USA (pp. 197–215). Berlin: Springer.

148

K. N. Mishra

Buyya, R., Yeo, C. S., & Venugopal, S. (2008). Market oriented Cloud Computing: Vision, hype, and reality for delivering IT services as computing utilities. In Proceedings of the 10th IEEE international conference on High Performance Computing and Communications (HPCC 2008, IEEE CS Press, Los Alamitos, CA, USA), Dalian, China, September 25–27, pp. 1–9. Caballero, J., Yin, H., Liang, Z., & Song, D. (2007). Polyglot: Automatic extraction of protocol message format using dynamic binary analysis. In P. Ning (Ed.), Proceedings of the 14th ACM conference on Computer and Communications Security – CCS 2007, held 28–31 October 2007 in Whistler, Canada (pp. 317–329). New York: ACM Publication. Chadwick, D. W., & Fatema, K. (2012). A privacy preserving authorization system for the Cloud. Journal of Computer and System Sciences, 78(5), 1359–1373. Chen, D., & Zhao, H. (2012). Data security and privacy protection issues in cloud computing. In Proceeding of the International Conference on Computer Science and Electronics Engineering (ICCSEE’12), pp. 1–5. Handley, M., & Rescorla, E. (2006). Internet denial-of-service considerations. In Internet engineering task force (Technical report, pp. 1–38). Hogan, M., Liu, F., Sokol, A., & Tong, J.  (2011). NIST cloud computing standards roadmap  – Version 1.0 (National Institute of Standard Technology Special Publication, 500-291, pp. 1–63). http://cloudcomputing.sys-con.com/node/612375/print. Accessed on May 5 2016. http://searchsecurity.techtarget.com/definition/authentication-authorization-and-accounting. Last Accessed On: May 5, 2017. Johnson, D.  H. (1999). The Insignificance of Statistical Significance Testing. The Journal of Wildlife Management, 63(3), 763–772. Kardaş, S., Çelik, S., Bingöl, M. A., & Levi, A. (2013). A new security and privacy framework for RFID in cloud computing. In Proceedings of the 5th IEEE international conference on cloud computing technology and science (CloudCom’13), pp. 1–5. Khalid, U., Ghafoor, A., Irum, M., & Awais Shibli, M. (2013). Cloud based secure and privacy enhanced authentication and authorization protocol. Procedia Computer Science, 22, 680–688. Khan, S., Shakil, K. A., & Alam, M. (2017a). Cloud based big data analytics: A survey of current research and future directions. In Big data analytics, electronic (pp. 629–640). Springer. Khan, S., Liub, X., Shakil, K. A., & Alam, M. (2017b). A survey on scholarly data: From big data perspective. Information Processing & Management, 53(4), 923–944. Khodadadi, F., Calheiros, R. N., & Buyya, R. (2015). A data-centric framework for development and deployment of internet of things applications in clouds. In Proceedings of the 10th IEEE international conference on Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP 2015), Singapore, April 7–9, pp. 1–6. Klein, D. A. (2013). Data security for digital data storage. U.S. Patent Application 14/022,095, 2013. Mahmood, Z. (2011). Data location and security issues in cloud computing. In Proceedings of the 2nd IEEE international conference on Emerging Intelligent Data and Web Technologies (EIDWT’11), pp. 49–54. Malhotra, S., Doja, M. N., Alam, B., & Alam, M. (2017). Generalized query processing mechanism in cloud database management system, big data analytics. Electronic, Springer, pp. 641–648. Masson, M. E. J., & Loftus, G. R. (2003). Using confidence intervals for graphically based data interpretation. Canadian Journal of Experimental Psychology, 57(3), 203. Mell, P., & Grance, T. (2009). The NIST definition of cloud computing, version 15, National Institute of standards and Technology (NIST), Information Technology Laboratory, pp. 1–3. Online Available On: www.csrc.nist.gov. Last Accessed on July 21, 2017. Miller, M. (2008). Cloud computing: Web based applications that change the way you work and collaborate online. Indianapolis: Que Publication. NIST Publication. (2001). Statistical test suite for random and pseudorandom number generators for cryptographic applications. http://csrc.nist.gov/publications/nistpubs/800-22/sp-800-22051501.pdf. Online available May 6, 2017.

7  Supervising Data Transmission Services Using Secure Cloud Based Validation…

149

Pandey, A., Tugnayat, R. M., & Tiwari, A. K. (2013). Data security framework for cloud computing networks. International Journal of Computer Engineering & Technology, 4(1), 178–181. Plummer, D. C., Smith, D., Bittman, T. J., Cearley, D. W., Cappuccio, D. J., Scott, D., Kumar, R., & Robertson, B. (2009). Gartner highlights five attributes of cloud computing. Gartner Report, G00167182, 1–5. Robert, A.  E., Manivasagam, G., Sasirekha, N., & Hemalatha, M. (2011). Reverse engineering for malicious code behaviour analysis using virtual security patching. International Journal of Computer Applications, 26(4), 41–45. Rogaway, P., & Black, J.  (2002). A block-cipher mode of operation for parallelizable message authentication. In L. R. Knudsen (Ed.), Proceedings of the international conference on theory and applications of cryptographic techniques, advances in cryptology – EUROCRYPT 2002, Lecture notes in computer science 2332, held 28 April–2 May 2002 in Amsterdam, Holland (pp 384–397). Berlin: Springer. Saldhana, A., Marian, R., Barbir, A., & Jabbar, S.  A. (2014). OASIS Cloud Authorization (CloudAuthZ). International Journal of Multimedia and Ubiquitous Engineering, 9(9), 81–90. Sedigh, A., Radhakrishnan, K., Campbell, C. E.-A., & Singh, D. (2014). Trust evaluation of the current security measures against key network attacks. MAGNT Research Report, 2(4), 161–171. Shakil, K. A., Sethi, S., & Alam, M. (2015). An effective framework for managing university data using a cloud based environment. computing for sustainable global development (INDIACom), IEEE 2nd international conference, pp. 1262–1266. Shakil, K. A., Anis, S., Khan, S., & Alam, M. (2017). Dengue disease prediction using weka data mining tool. In Proceedings of IIRAJ International Conference (ICCI-SEM-2K17). Singh, I., Rai, R., & Murarker, S. (2015a). Password authentication in cloud. International Journal of Engineering Research and Applications, 9(2 (Part I)), 56–59. Singh, I., Mishra, K. N., Alberti, A., Singh, D., & Jara, A. (2015b). A novel privacy and security network for the cloud network services. 17th IEEE international conference on advanced communication technology, pp. 355–359. Sun, D., Chang, G., Sun, L., & Wang, X. (2011). Surveying and analyzing security, privacy and trust issues in cloud computing environments. In Proceedings of the international conference on advanced in Control Engineering and Information Science (CEIS’11), pp. 1–5. Tan, Y., Sengupta, S., & Subbalakshmi, K.  P. (2011). Analysis of coordinated denial-of-service attacks in ieee 802.22 networks. Selected Areas in Communications, IEEE Journal, 29(4), 890–902. Van Bon, J., & Van Der Veen, A. (2007). Foundations of IT service management based on ITIL (Vol. 3). Zaltbommel: Van Haren Publishing. Whiting, D., Housley, R., & Ferguson, N. (2002). AES encryption & authentication using CTR mode & CBC-MAC. http://csrc.nist.gov/groups/ST/toolkit/BCM/current_modes.html#03. Online available May 6, 2017. Wijaya, C. (2011). Performance analysis of dynamic routing protocol EIGRP and OSPF in IPv4 and IPv6 network. First international conference on Informatics and Computational Intelligence (ICI), pp. 355–360. Younis, M. Y. A., & Kifayat, K. (2013). Secure cloud computing for critical infrastructure: A survey. UK Liverpool John Moores University.

Part III

IoT Challenges and Issues

Chapter 8

Tackling Jamming Attacks in IoT N. Ambika

Abstract  Internet-of-things has brought in solutions for many problems. These devices work on sensor technology. They aim in sensing and making suitable adjustments in environment to improve the environment. These unsupervised devices are liable to different kinds of attacks. The proposed paper provides solution for reactive jamming attack. The previous contribution aims at using artificial noise to distract the adversary and accomplishes the transmission task. The proposed work tries to bring in the jammer in confidence by using the cooperation of all the devices in the network. Comparing with the previous work the proposed work conserves 7.89% of energy, increases the confidence of jammer by5.66% and increases communication overhead by 7.46%. Keywords  Jamming attack · Internet of things · Cooperative strategy · Security

8.1  Introduction Internet-of-things (Alaba et al. 2017; Atzori et al. 2010; Ahmad 2014; Stankovic 2014) provides solutions for a lot of unattended problems. The technology aids the devices to communicate with each other irrespective of their capability and capacity. Many applications (Hegde and Kumar 2017; Liu and Zhu 2014) use this technology to bring in a better outcome. Some of the Smart homes (Al-Ali 2017; Alam et al. 2012; Mano et al. 2016) applications include aiding elderly monitoring (Yang et al. 2013; Rathore et al. 2016; Muhammad et al. 2017; Arcelus et al. 2007), to provide security (Andrea et  al. 2015; Jabarullah et  al. 2012; Alam et  al. 2014; Sivaraman et al. 2015), home automation (Kodali et al. 2016). Other areas where these devices are utilized include designing smart city (Mitton et al. 2012), smart parking system (Khanna and Anand 2016), smart transportation (Melis et al. 2016; Sutar et al. 2016; Masek et al. 2016; Jisha et al. 2017), smart hospital management (Yu et al. 2012), agriculture (Mat et al. 2016; Popović et al. 2017), smart industry (Shrouf et al. 2014). N. Ambika (*) Department of Computer Applications, SSMRV College, Bangalore, India © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_8

153

154

N. Ambika

These devices are programmed to do certain activities to minimize human efforts. Human intervention would be only at the time of setting up the application, performing some kind of activation and removing the setup. Hence most of the time these devices are left unsupervised. The adversary will be able to introduce any kind of attack in these environments. Jamming attack (Grover et al. 2014; Babar et al. 2013; Xu et al. 2005) is one such attack which exhibits behavior with some variations based on its type – reactive jamming attack (Wilhelm et al. 2011), constant jammers (Han et al. 2018), Deceptive jammers (Lee et al. 2014), Random jammers (Wei et al. 2014; Sanguanpong 2018), Statistical jammer (Ren and Wang 2013) and Protocol-aware jammer (Sufyan et al. 2013; Toledo and Wang 2008). In this proposal, the reactive jamming attack is considered and a novel method is suggested to tackle the same. The proposed work aims – • Gaining the cooperation of the nodes of the network. • Creating real-time illusion to the adversary. • Tackling the attack better than the previous work (Pang and Xue 2017). • The work is divided into 6 sections. Literature survey is detailed in Sect. 8.2. The proposed work is explained in Sect. 8.3. The analysis of the work is given in Sect. 8.4. Simulation of the work is detailed in the next section. The work is concluded in final section.

8.2  Literature Survey Internet-of-things is used to link two devices and enables communication among them. As these devices are not under continuous supervision they are liable to different kinds of attacks. The proposed work aids in minimizing jamming attack in the environment. The reactive jamming attack is considered in the work. Radio interference attacks are responsible to cause serious injury in the network. The following section provides a detailed study of the jamming attacks proposed by various authors. In Xu et al. (2005) jamming attacks are analyzed in wireless networks. The work is evaluated in two stages. In the first stage, the problems that crop up when radio interference attack is encountered is scrutinized. In the second stage, critical issues of diagnosing the presence of the attack are dissected. Four kinds of jamming attacks are introduced into the network and the effect on the wireless network, their ability to receive and send messages is evaluated. The work is analyzed considering the signal strength and carrier sensing time. The proposed work used Mica2 as the platform to conduct the respective experiments. The MAC protocol (Hamza et al. 2016) waits for the idle slot for a random amount of time to transmit the packets. The four kinds of jammer – constant jammer, reactive jammer, random jammer, and deceptive jammer were evaluated. Berkeley motes are used in the system. They useChipCon CC1000 RF transceiver and TinyOS operating system were used to

8  Tackling Jamming Attacks in IoT

155

implement and evaluate the working of the study under the four kinds of jamming (Noubir et al. 2011) attacks. The channels were disabled to sense the packets and the backoff operations to bypass MAC protocol to introduce the attacks in the network. The study will not be able to analyze the difference between the congested packet and jamming attack. Reactive jamming is evaluated in Wilhelm et al. (2011). An IEEE 802.15.4 network is used to examine the proposed work. The loss occurred at the physical layer are identified and a solution is provided to tackle reactive jamming attack. To discover the attack detection of jamming, scheduling and initialization of jamming and jamming burst to destroy the packet is calculated against the minimal time required for jamming. The proposed work concluded that minimal jamming duration is 26 μs, jamming initialization time was set to 15 μs and the detector added a delay of less than 4 μs. The accounted data is used to check the same against USRP2-based reactive jamming prototype. In Raya and Hubaux (2007) jamming attack was experimented on vehicular networks. Analyzing the threat analysis appropriate security architecture was devised. Digital signatures were used to protect the message legacy from the adversaries. The work adopts to use a set of public/private keys pairs to sign the messages digitally and validate (Shakil et al. 2017) themselves to the receivers. To tackle revocation of certificates a set of revocation protocols are used. Revocation Protocol of the Tamper-proof Device (RTPD), Revocation protocol using Compressed Certificate Revocation list (RCCRL) and Distributed Revocation Protocol (DRP) are utilized in the work. Using the protocol the malicious behavior can be detected. A game-theoretic analysis is used to avoid the attack of an airborne jammer on a communication channel (Bhattacharya and Başar 2010). The kinematic model of the nodes is considered. To obtain optimal strategies, saddle point strategies are used. This is a zero-sum game which has a set of strategies for the players are used to arrive at the equations governing the saddle point strategies. In Brown et  al. (2006) the authors have considered an attacker disrupting an encrypted victim wireless Adhoc network using jamming practice. Jamming at the Transport/network layer is considered in the work. The analysis revealed that TCP-SYN and TCP-­ SYN-­ ACK packets prevent TCP connection. ARP-REQUEST and ARQ-­ RESPONSE will prevent IP from associating IP and MAC address. Sensing is considered to be both online and offline. Online sensing packets are identified as they are received by the destination. Offline sensing is used to categorize the packets received in the history doings. The spread spectrum proposed in the work provides a jamming immunity proportional to the spreading factor. The authors used Linux laptops running AODV-UU protocol while conducting testbed. The authors in Li et al. (2007) have provided a solution for single-channel wireless sensor networks. The sophisticated jammer in the work controls the probability of the jamming and transmission range aims to corrupt the communication links in the network. The monitoring node is responsible to monitor the doings of the other nodes in the network. The supervising node conveys the alert message to the respective. The monitoring node is responsible to estimate the percentage of incurred collisions and draw out the outcome of optimal detection test. The network computes

156

N. Ambika

channel access probability to aid in minimizing jamming detection and reporting time. The study focuses to understand the problems regarding the structure, identifying the tradeoffs and capturing the impact of different parameters on its performance. The study is made to bring in a strategy which aims to handle the situation in case of a knowledge-based/without knowledge-based scenario. The network and monitor node without the knowledge uses the formulation and resolving optimization problems. Path-based Dos attack is dealt in Deng et al. (2005). The wireless sensor network will undergo serious damage. The intruder either injects replays the packets into the network or spurious packets. The adversary aims to block the path by introducing jamming packets and initiating the base station to re-boot the OHC (a sequence of numbers) for all the nodes on the path. Two methodologies are adapted aiding to minimize jamming and forgo the above scenario. OHC proactive bootstrapping employs to bootstrap the nodes near the jamming path and refreshing the OHC sequence. Another approach is known as lazy OHC bootstrapping where the addition of a new node to the network does not commence bootstrapping but still maintains resilience to stop Dos attack. In Hamieh and Ben-Othman (2009) the authors have designed a solution for jamming attacks in mobile Adhoc networks. The authors have used a scenario where the jammer gets itself into its doings when valid radioactivity is signaled. The adversary moves to sleep state when it is not jamming the signals. To find the solution to this kind of attack, dependence among the periods of error and correct reception times are analyzed. The correlation coefficient is used to compute the same. The coefficient is the statistical measure of the relation between two random variables. To evaluate the jammer, error probability and correlation coefficient is measured. In Sampath et al. (2007) 802.11 networks are considered. A detailed study on Single-channel jamming and multi-channel jamming attacks is done in this work. In the case of single-channel jamming attack, the adversary transmits high power signals on the channel. The impact of jamming depends on the size of the packets, length of inter-jamming interval and size of jamming packets. The adversary is observed to transmit 50-byte size packets to disrupt the data link. According to the experiment conducted in the work 100%,—40% of packets suffer from degradation. Considering the second scenario, multiple radio devices are used to cause multi-­ channel jamming attacks. The impact can cause 5-100 ms switching delay. The solution to Spoofing-jamming attack in wireless smart grid network is detailed in Gai et  al. (2017). The work considers both jamming and spoofing to intrude in the communication. The frequencies used to transmit jamming packets vary with time. Spectrum adversarial occupations are used to maximize the attack effect. The study uses dynamic programming to solve power distribution optimal problem. The study is evaluated using the simulator MAS-SIM. In Pang and Xue (2017) the authors have provided a solution to tackle intelligent jamming attacks. The method uses the adversary location as a precondition to restoring the transmission links. The synthetic noise generated by synergistic sensor nodes the channel of the adversary is degraded. New types of a jamming attack like clear-to-send frame attack, DATA frame attack and ACK frame attack is considered

8  Tackling Jamming Attacks in IoT

157

to provide a solution. The anti-jamming method can tackle the adversary in carrier sense multiple access with collision avoidance protocol (CSMA/CA). An artificial noise is created and the same use jammer localization methods as a precondition. A cooperative strategy that aids in jamming the intelligent adversary is enabled at the physical layer. The high signal-to-noise ratio disables the jammer to decipher the receiving packets. This methodology aids the network not to interfere in a special frame timely. The ratio of the packet transmitted and received are calculated in the work. The proposed work uses a similar scenario with some modifications is utilized in IoT (Riedel et al. 2010; Lopez et al. 2017). The proposal aids to minimize energy (Chiu et al. 2017) in the devices.

8.3  Proposed Study Reactive jamming attack (Fadele et al. 2018; Sciancalepore et al. 2018) the adversary emits radio signals on sensing the activity in the channel. These packets keep the channel busy. The packets dispatched by the legitimate devices may die before reaching the destination or they may get lost in the channel. In Pang and Xue (2017) the authors have provided a solution to tackle intelligent jamming attacks. The method uses the jammer location as a prerequisite to restoring the communication links. Based on using the artificial noise generated by synergistic sensor nodes (Narayanan et  al. 2016) the channel of the jammer is degraded. New types of a jamming attack like clear to send frame jammer, DATA frame jammer and ACK frame jammer is considered to provide a solution. The anti-jamming method can tackle the jammer in carrier sense multiple access with collision avoidance protocol (CSMA/CA). An artificial noise is created and the same use jammer localization methods as a precondition. A cooperative strategy that aids in jamming the intelligent jammer is enabled at the physical layer. The high signal-to-noise ratio disables the jammer to decode the receiving packets. This methodology aids the network not to interfere in a special frame timely. Packet-send-ratio (PSR) and packet delivery ratio (PDR) is calculated in the work. The proposed work uses a similar scenario with some modifications is utilized in IoT (Mala et al. 2017). The proposal aids to minimize energy in the devices. In the proposed work the devices cooperate to create a different illusion to the adversary. The adversary will try engaging its packets in the different direction and hence the actual packets to be successfully delivered to the respective destination.

8.3.1  Notations Used in the Study The notations and its decription is given in Table 8.1.

158

N. Ambika

Table 8.1  Notations and its description

Notations N Hello Li Ack idi

Description Network in consideration Hello message Location of ith device Acknowledgement message Identification of ith device Null packet with identification of ith device

8.3.2  Assumptions Made in the Study • The network is free of all kinds of attacks till the initial identification of communicating device is made. • The network is liable for reactive jamming attack. • The system uses carrier sense multiple access with collision avoidance protocol (CSMA/CA).

8.3.3  Creating Awareness The IoT devices first communicate among themselves to know each other. They send their location and their identity during communication. In notation (1) let A is broadcasting the hello message along with its location and identification to the network. Let r be the communication radius. The devices which are able to hear each other will respond with the Ack message. The acknowledgement message contains Ack, location and identification of the sender. In the notation (2) IoT device B is transmitting the message to the device A.

A ® ( hello, Li , idi ) : N





B ® ( Ack, L j , id j ) : A



(8.1) (8.2)

8.3.4  Identifying the Jammer Location The devices irrespective of the type have a certain capacity to transmit the packets. The number of packets for a device varies but it is accountable. The devices maintain the packet delivery ratio (PDR) and the packet sent ratio (PSR) entries. The regular report of the packets delivered and received is transmitted to the control unit. Analyzing the device capacity, the frequency, threshold with device id for each device is set. The same is communicated by the control unit to the devices. If the

8  Tackling Jamming Attacks in IoT

159

Fig. 8.1  Representation of the network of IoT devices with an adversary

devices receive data from inappropriate source, the same is reported to the control unit. Other parameters – packets received, frequency are reported of the inappropriate device to the control unit. The control unit analyses the same and if the device id is not registered in its storage, the jammer is identified and notified to the network.

8.3.5  Choosing Appropriate Path for Delivery From the Fig. 8.1, Let us assume that the device C has to deliver the data to the device F. as the jammer gets notified, the distraction approach is adopted. The device C sends a null packet to D with the identification of device A. in the Eq. (8.3) the device C is delivering the null packet to the device D. Once the device D gets the message, the device D sends some packets to A to provide a distraction. The jammer tries to introduce the packets between D and A. In the meantime, device C delivers the packet to F.

C ® null, id A : D

(8.3)

8.4  Analysis of the Work Security (Farash et al. 2016) is necessary in the network to increase reliability (Li et  al. 2012a, b), confidentiality and privacy. In Pang and Xue (2017) the authors have provided a solution to tackle intelligent jamming attacks. The method uses the

160

N. Ambika

adversary location as a precondition to restoring the transmission links. The synthetic noise generated by synergistic sensor nodes the channel of the adversary is degraded. New types of a jamming attack like clear-to-send frame attack, DATA frame attack and ACK frame attack is considered to provide a solution. The anti-­ jamming method can tackle the adversary in carrier sense multiple access with collision avoidance protocol (CSMA/CA). An artificial noise is created and the same use jammer localization methods as a precondition. A cooperative strategy that aids in jamming the intelligent adversary is enabled at the physical layer. The high signal-­to-noise ratio disables the jammer to decipher the receiving packets. This methodology aids the network not to interfere in a special frame timely. The ratio of the packet transmitted and received are calculated in the work. The proposed work uses a similar scenario with some modifications is utilized in IoT. The proposal aids to minimize energy in the devices.

8.4.1  Energy Consumption Energy is one of the important resources. As these systems are installed in a hospital or home setup, tackling the attacks becomes essential. In the proposed work the system uses carrier sense multiple access with collision avoidance protocol (CSMA/CA). The devices cooperate among themselves for effective communication. In Pang and Xue (2017) artificial noise is generated. When the noise is generated at the same location every time, the jammer will be able to identify the same and try performing using a different strategy. To avoid such a scenario, the proposed work has suggested some real-time methodology. The method aids in confusing the jammer with its tactics. The work (Pang and Xue 2017) has to use a lot of energy to create artificial noise. Comparing the proposed work with the work (Pang and Xue 2017), the proposed work uses 7.89% of less energy and gains the confidence of the jammer. The same is depicted in Fig. 8.2.

8.4.2  Taking Jammer into Confidence The work (Pang and Xue 2017) uses artificial noise to distract the jammer. The proposed work is also accomplishing the task by the cooperation of the devices in the network. The jammer position is identified in the proposed work using data provided by the control unit. When the devices have to send some communication to the other device, to distract the jammer the device transmits a null packet to another device. The device that receives the packet sends some packets to the mentioned device. In the meantime, the actual packets reach the destination without any hindrance. Hence the proposed work proves to be 5.66% more reliable (Ahmad 2014) than the work (Pang and Xue 2017). The same is represented in Fig. 8.3.

8  Tackling Jamming Attacks in IoT

161

Fig. 8.2  Comparison of Energy consumption (Joules)

Fig. 8.3  Comparison of confidence gain of the jammer

Fig. 8.4  Comparison of the work (Pang and Xue 2017) with the proposed work w.r.t communication overhead

3000 2500 2000

Proposed work

1500

(Pang & Xue, 2017)

1000 500 0 1

2

3

4

5

6

8.4.3  Communication Overhead The work (Pang and Xue 2017) does not impose any load on the devices rather the noise is generated by the artificial noise generated meant for the same purpose. The proposed work adds some amount of transmission load on the other devices to accomplish the task. Comparing the work (Pang and Xue 2017) with the proposed work, the proposed work has 7.46% overhead. The same is depicted in Fig. 8.4.

8.4.4  Simulation of the Work The work is simulated using NS2. The following parameters as described in the Table 8.2 are used to simulate the work.

162 Table 8.2  Parameters used in the simulation work

N. Ambika Parameters Dimension No of devices considered Jammer considered Number of packets delivered per unit time Simulation time Null packet size Packet size identification of device Packet size of the location Packet size of hello/ acknowledgement packet

Description 200 m ∗ 200 m 5 1 248 60s 13 bits 12 bits 24 bits 6 bits

8.5  Conclusion Internet-of-things is a boon to mankind. The technology aids in bringing different kinds of devices with varying size and capacity into a common platform for communication. As these devices are unsupervised the network is liable for different kinds of attacks. Reactive jammer attack is one such attack where the adversary introduces the attack on sensing any packets in the channel. In this scenario, the source will not be successfully transmitting all its packets to the desired destination. In the earlier work proposed by Pang and Xue, artificial noise was created to distract the adversary from the desired location. The proposed work suggested using the cooperation of the devices to bring in the real-time scenario and gain the confidence of the adversary. The proposed work increases its confidence with an adversary by 5.66% compared to work proposed by Pang & Xue and energy is conserved by 7.89%. The communication overhead is increased by 7.46% compared with the previous work.

References Ahmad, M. (2014). Reliability models for the Internet of Things: A paradigm shift. In IEEE international symposium on software reliability engineering workshops, Naples, Italy. Alaba, F. A., Othman, M., Hashem, I. A. T., & Alotaibi, F. (2017). Internet of Things security: A survey. Journal of Network and Computer Applications, 10–28. Al-Ali, et al. (2017). A smart home energy management system using IoT and big data analytics approach. IEEE Transactions on Consumer Electronics, 63(4), 426–434. Alam, M. R., Reaz, M. B. I., & Ali, M. M. (2012). SPEED: An inhabitant activity prediction algorithm for smart homes. IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans, 42, 985–990. Alam, M., Shakil, K.  A., Javed, M.  S., Ansari, M., & Ambreen. (2014). Detect and filter traffic attack through cloud trace back and neural network. In International Conference of Data Mining and Knowledge Engineering ( ICDMKE). London, UK: Imperial College.

8  Tackling Jamming Attacks in IoT

163

Andrea, I., Chrysostomou, C., & Hadjichristofi, G. (2015). Internet of Things: Security vulnerabilities and challenges. In IEEE Symposium on Computers and Communication (ISCC), IEEE (pp. 180–187). Arcelus, A., Jones, M. H., Goubran, R., & Knoefel, F. (2007). Integration of smart home technologies in a health monitoring system for the elderly. In 21st International Conference on Advanced Information Networking and Applications Workshops (AINAW'07), Niagara Falls, ON, Canada (Vol. 2, pp. 820–825). Atzori, L., Iera, A., & Morabito, G. (2010). The Internet of Things: A survey. Computer Networks, 54, 2787–2805. Babar, S.  D., Prasad, N.  R., & Prasad, R. (2013). Jamming attack: Behavioral modelling and analysis. In International Conference on Wireless Communications, Vehicular Technology, Information Theory and Aerospace & Electronic Systems (VITAE) (pp. 1–5). Bhattacharya, S., & Başar, T. (2010). Game-theoretic analysis of an aerial jamming attack on a UAV communication network. In American control conference, Baltimore, MD, USA (pp. 818–823). Brown, T. X., James, J. E., & Sethi, A. (2006). Jamming and sensing of encrypted wireless ad hoc networks. In 7th ACM international symposium on Mobile ad hoc networking and computing, Florence, Italy (pp. 120–130). Chiu, W.-Y., Sun, H., Thompson, J., Nakayama, K., & Zhang, S. (2017). IoT and information processing in smart energy applications. IEEE Communications Magazine, 55, 44. Deng, J., Han, R., & Mishra, S. (2005). Defending against path-based DoS attacks in wireless sensor networks. In 3rd ACM workshop on Security of ad hoc and sensor networks , Alexandria, VA, USA (pp. 89–96). Fadele, A. A., Othman, M., Hashem, I. A. T., Yaqoob, I., Imran, M., & Shoaib, M. (2018). A novel countermeasure technique for reactive jamming attack in internet of things. Multimedia Tools and Applications, 1–22. Farash, M. S., Kumari, S., & Bakhtiari, M. (2016). Cryptanalysis and improvement of a robust smart card secured authentication scheme on SIP using elliptic curve cryptography. Multimedia Tools and Applications, 75(8), 4485–4504. Gai, K., Qiu, M., Ming, Z., Zhao, H., & Qiu, L. (2017). Spoofing-jamming attack strategy using optimal power distributions in wireless smart grid networks. IEEE Transactions on Smart Grid, 2431–2439. Grover, K., Lim, A., & Yang, Q. (2014). Jamming and anti-jamming techniques in wireless networks: A survey. International Journal of Ad Hoc and Ubiquitous Computing, 197–215. Hamieh, A., & Ben-Othman, J. (2009). Detection of jamming attacks in wireless ad hoc networks using error distribution. In IEEE international conference on communications, Dresden, Germany (pp. 1–6). Hamza, T., Kaddoum, G., Meddeb, A., & Matar, G. (2016). A survey on intelligent MAC layer jamming attacks and countermeasures in WSNs. In 84th vehicular technology conference (pp. 1–5). Han, G., Liu, L., Zhang, W., & Chan, S. (2018). A hierarchical jammed-area mapping service for ubiquitous communication in smart communities. IEEE Communications Magazine, 92–98. Hegde, S., & Kumar, S. G. (2017). IoT approach to save life using GPS for the traveller during accident. In IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI), Chennai, India (pp. 2424–2428). Jabarullah, B. M., Saxena, S., Kennedy Babu, C. N., & Alam, M. (2012). Hybrid approach of face recognition. Cyber Times International Journal of Technology & Management, 6(1), 6–12. Jisha, R. C., Jyothindranath, A., & Kumary, L. S. (2017). Iot based school bus tracking and arrival time prediction. In International Conference on Advances in Computing, Communications and Informatics (ICACCI), Udupi, India. Khanna, A., & Anand, R. (2016). IoT based smart parking system. In International Conference on Internet of Things and Applications (IOTA), Pune, India (pp. 266–270).

164

N. Ambika

Kodali, R. K., Jain, V., Bose, S., & Boppana, L. (2016). IoT based smart security and home automation system. In International conference on computing, communication and automation (ICCCA), Noida, India (pp. 1286–1289). Lee, T. H., Wen, C. H., Chang, L. H., Chiang, H. S., & Hsieh, M. C. (2014). A lightweight intrusion detection scheme based on energy consumption analysis in 6LowPAN. Advanced Technologies, Embedded and Multimedia for Human-Centric Computing, 1205–1213. Li, M., Koutsopoulos, I., & Poovendran, R. (2007). Optimal jamming attacks and network defense policies in wireless sensor networks. In 26th IEEE international conference on computer communications, Barcelona, Spain (pp. 1307–1315). Li, L., Jin, Z., Li, G., Zheng, L., & Wei, Q. (2012a). Modeling and analyzing the reliability and cost of service composition in the IoT: A probabilistic approach. In IEEE 19th International Conference on In Web Services (pp. 584–591). Li, W., Miao, Y., Tang, Y. W., Liu, D., & Hu, B. (2012b). IOT system reliability testing and evaluate technology. Software, 1. Liu, S. J., & Zhu, G. Q. (2014). The application of GIS and IOT technology on building fire evacuation. Procedia Engineering, 71, 577–582. Lopez, D. A. R., Lopez, J. R. R., Prieto, M. A. Z., & Quinde, L. D. S. (2017). Towards a method for the integration of IoT and GIS applications deployed on cloud platforms. In International Conference on Information Systems and Computer Science (INCISCOS), Quito, Ecuador. Mala, N. S., Thushara, S. S., & Subbiah, S. (2017). Navigation gadget for visually impaired based on IoT.  In 2nd International Conference on Computing and Communications Technologies (ICCCT), Chennai, India. Mano, L. Y., Faiçal, B. S., Nakamura, L. H., Gomes, P. H., Libralon, G. L., Meneguete, R. I., & Ueyama, J. (2016). Exploiting IoT technologies for enhancing Health Smart Homes through patient identification and emotion recognition. Computer Communications, 89, 178–190. Masek, P., et al. (2016). A harmonized perspective on transportation management in Smart Cities: The novel IoT-driven environment for road traffic modeling. Sensors, 16(11), 1872. Mat, I., Kassim, M. R. M., Harun, A. N., & Yusoff, I. M. (2016). IoT in precision agriculture applications using wireless moisture sensor network. In IEEE Conference on Open Systems (ICOS), Langkawi, Malaysia (pp. 24–29). Melis, A., Prandini, M., Sartori, L., & Callegati, F. (2016). Public transportation, IoT, trust and urban habits. In International conference on internet science (pp. 318–325). Mitton, N., Papavassiliou, S., Puliafito, A., & Trivedi, K. S. (2012). Combining cloud and sensors in a smart city environment. EURASIP Journal on Wireless Communications and Networking, 1–10. Muhammad, G., Rahman, S.  M. M., Alelaiwi, A., & Alamri, A. (2017). Smart health solution integrating IoT and cloud: A case study of voice pathology monitoring. IEEE Communications Magazine, 55(1), 69–73. Narayanan, R. P., Sarath, T. V., & Vineeth, V. V. (2016). Survey on motes used in wireless sensor networks: Performance & Parametric Analysis. Wireless Sensor Network, 08, 51–60. Noubir, G., Rajaraman, R., Sheng, B., & Thapa, B. (2011). On the robustness of IEEE 802.11 rate adaptation algorithms against smart jamming. In Fourth ACM conference on Wireless network security (pp. 97–108). Pang, L., & Xue, Z. (2017). A novel anti-jamming method in wireless sensor networks: Using artificial noise to actively interfere the intelligent jammer. In 4th international conference on systems and informatics (pp. 954–959). Popović, T., Latinović, N., Pešić, A., Zečević, Ž., Krstajić, B., & Djukanović, S. (2017). Architecting an IoT-enabled platform for precision agriculture and ecological monitoring: A case study. Computers and Electronics in Agriculture, 140, 255–265. Rathore, M. M., Ahmad, A., Paul, A., Wan, J., & Zhang, D. (2016). Real-time medical emergency response system: Exploiting IoT and big data for public health. Journal of Medical Systems, 40(12), 283.

8  Tackling Jamming Attacks in IoT

165

Raya, M., & Hubaux, J.  P. (2007). Securing vehicular ad hoc networks. Journal of Computer Security, 15, 39–68. Ren, K., & Wang, Q. (2013). Opportunistic spectrum access: From stochastic channels to non-­ stochastic channels. IEEE Wireless Communications, 128–135. Riedel, T., Fantana, N., Genaid, A., Yordanov, D., Schmidtke, H. R., & Beigl, M. (2010). Using web service gateways and code generation for sustainable IoT system development. Internet of Things, 1–8. Sampath, A., Dai, H., Zheng, H., & Zhao, B.  Y. (2007). Multi-channel jamming attacks using cognitive radios. In 16th international conference on computer communications and networks, Honolulu, HI, USA (pp. 352–357). Sanguanpong, S. (2018). Secrecy outage performance analysis for energy harvesting sensor networks with a jammer using relay selection strategy. In Special section on security and trusted computing for industrial Internet of Things (pp. 23406–23419). Sciancalepore, S., Oligeri, G., & Di Pietro, R. (2018). Strength of Crowd (SOC)—Defeating a reactive jammer in IoT with decoy messages. Sensors, 18, 3492. Shakil, K. A., Zareen, F. J., Alam, M., & Jabin, S. (2017). BAMHealthCloud: A biometric authentication and data management system for healthcare data in cloud. Journal of King Saud University – Computer and Information Sciences. Shrouf, F., Ordieres, J., & Miragliotta, G. (2014). Smart factories in industry 4.0: A review of the concept and of energy management approached in production based on the Internet of Things paradigm. In IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), Bandar Sunway, Malaysia (pp. 697–701). Sivaraman, V., Gharakheili, H. H., Vishwanath, A., Boreli, R., & Mehani, O. (2015). Network-­ level security and privacy control for smart-home IoT devices. In IEEE 11th International conference on wireless and mobile computing, networking and communications (WiMob), Abu Dhabi, United Arab Emirates (pp. 163–167). Stankovic, J. A. (2014). Research directions for the Internet of Things. IEEE Internet of Things Journal, 1(1), 3–9. Sufyan, N., Saqib, N. A., & Zia, M. (2013). Detection of jamming attacks in 802.11 b wireless networks. EURASIP Journal on Wireless Communications and Networking, 208. Sutar, S. H., Koul, R., & Suryavanshi, R. (2016). Integration of Smart Phone and IOT for development of smart public transportation system. In International Conference on Internet of Things and Applications (IOTA), Pune, India (pp. 73–78). Toledo, A.  L., & Wang, X. (2008). Robust detection of MAC layer denial-of-service attacks in CSMA/CA wireless networks. IEEE Transactions on Information Forensics and Security, 347–358. Wei, X., Hu, Y., Fan, J., & Kan, B. (2014). A jammer deployment method for multi-hop wireless network based on degree distribution. In Ninth International Conference on Broadband and Wireless Computing, Communication and Applications (BWCCA) (pp. 261–266). Wilhelm, M., Martinovic, I., Schmitt, J. B., & Lenders, V. (2011). Short paper: Reactive jamming in wireless networks: How realistic is the threat? In Fourth ACM conference on Wireless network security, Hamburg, Germany (pp. 47–52). Xu, W., Trappe, W., Zhang, Y., & Wood, T. (2005). The feasibility of launching and detecting jamming attacks in wireless networks. In Proceedings of the 6th ACM international symposium on Mobile ad hoc networking and computing (pp. 46–57). ACM. Yang, L., Yang, S. H., & Plotnick, L. (2013). How the Internet of Things technology enhances emergency response operations. Technological Forecasting and Social Change, 80(9), 1854–1867. Yu, L., Lu, Y., & Zhu, X. (2012). Smart hospital based on Internet of Things. Journal of Networks, 7(10), 1654.

Chapter 9

Bioinspired Techniques for Data Security in IoT S. R. Mani Sekhar, G. M. Siddesh, Anjaneya Tiwari, and Ankit Anand

Abstract  Data Security is a protective privacy measure which is used for prevention of unauthorized access in various domains such as computers network, web application, database etc. Bio-inspired computing is a branch of Machine learning deals with the biological properties of living organism. These techniques are used in the fields of data security and feature extraction in combination with different IoT architecture for secure data transition and maintenance. This chapter discusses some of the biologically inspired computational algorithms like Ant Colony Optimization, Artificial Bee Colony, Genetic bee colony, and Firefly which are used in solving real world problem in the field of IoT for providing data security. Finally the chapter concludes by stating the effectiveness of these algorithms on the bases of the case studies and thus provides a wider perspective towards Bio-inspired techniques and their use for data security in IoT. Keywords  Security · Bioinspired · IoT · Internet of Things · Ant colony optimization · Artificial bee colony optimization · Genetic bee colony · Firefly

9.1  Introduction Data Security are protective privacy measures which are used for prevention of unauthorized access in various domains such as computers, websites or databases. It is also used for the protection of data from getting corrupted. Since data security is necessary in the field of information technology, it is also known as information security or computer security. It is an important aspect of IT in the organisations of all sizes and types. Authentication is the most common method of applying data security with the help of password, usernames or some other kind of data such as S. R. Mani Sekhar (*) · G. M. Siddesh · A. Tiwari · A. Anand Department of Information Science & Engineering, Ramaiah Institute of Technology, Bangalore, Karnataka, India © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_9

167

168

S. R. Mani Sekhar et al.

biometric in order to verify someone’s identity before granting them access to ­system. Some examples of data security in day-to-day world are masking of data and data erasure. Encryption is a data security technology where digital data, hard drives, software or hardware are encrypted. Some of the very important fields where data security is being applied include health care record security, protection of a brand, capital and customer’s information and in IoT applications. Data security includes a set of standards which helps in the protection of data from getting destructed accidentally or intentionally. Methods such as applying data security using administrative controls, organisational standards, logical controls, physical security or other safeguarding techniques helps in limiting access to malicious or unauthorized users or processes. The main objective of using Biological techniques is to incorporate how living being solve their daily life problems by applying various characteristic functions and use these in solving complex computerised problems. The foundations on which the biologically inspired computational algorithms depends upon, includes subjects such as biology, computer science and mathematics. The analysis of these fields in relation with nature allows the creation of biologically inspired algorithms. Bio inspired computing deals with studying various designs, patterns and variations among the living beings which are further used for computing different compound problems such as the NP hard problems etc. These techniques are used in the fields of data security and feature extraction in combination with its integration with IoT for providing data security (Banu et al. 2016; Sari 2017). In the following section, this chapter discusses about key objectives of data security in IoT and how it can be achieved using Bio inspired algorithms. Firstly, it gives a detailed information about what the keyword actually means like data security in IoT, what is Bio inspired computing, the various types of algorithms available. After this, the chapter mainly describes three algorithms which are currently in use for securing data and proved to very efficient. The algorithms which are covered in this chapter include Ant colony optimization algorithm, artificial bee colony optimization algorithm, genetic bee colony algorithm, and firefly algorithm. In the detailed description of each algorithm the chapter presents two case studies in each algorithm to widen the knowledge base about these algorithms thus providing better understanding. After this we sum up by stating the effectiveness of these algorithms and by providing a detailed conclusion about the chapter.

9.2  Data Security in IoT IoT refers to internet of things which is a combination of network of various devices and appliances containing software, electronics and connectivity which help in the connection and interaction of data along with exchange of data. IoT includes involvement of devices beyond that of standard devices such as mobiles, laptops or

9  Bioinspired Techniques for Data Security in IoT

169

pc’s. IoT allows interaction of these devices which communicate and interact with each other via internet thus allows easy controlling and monitoring of these devices. IoT is seen as next revolution in the coming future so it is very important that protection of data in IoT technology is assured in order to allow fluent communication between devices with very little or no human intervention. IoT will include integration of communication devices, chips and sensors along with appliances (physical objects) which will enable communication between them and other devices including laptops, smart phones, computers and cloud servers. IoT has certain implications which must be considered including: • Generation of huge amount of data by these devices, so faster networks, large storage and wider bandwidth are needed to support its growth. • Requirement of an ecosystem for hosting of these devices to make an interoperable system. • Switching from IPv4 to IPv6 which is a more scalable protocol. Measures such as well protected devices, inculcating best practices for protection against security exploits and an interoperable interface are very important before switching to IoT.

9.3  Bioinspired Computing It is a research based method which is aimed at giving solutions to problems with the help of computer models, involving biological principles in the nature. This approach puts focus on dependability and is a bottom-up approach. The three key features of this approach are connectionism, emergence and social behavior. It puts the use of computers in modeling the living phenomena and studying of life simultaneously in order to improve its usage. The models of computers derive their abstractions from living organism of real world along with their behavior socially. This field is a branch of machine learning relating very closely to artificial intelligence. Since it involves subjects such as computer science, mathematics and biology, it can go up to a wide level in studying the diverse patterns and variations found among the living organisms thereby helping in building a better computer system for solving complex problems in future. A few of the most popular bio-inspired computations algorithms which are used for creating a secured framework in the field of data security are neural networks, particle swarm algorithm, genetic algorithm, bee colony algorithm, ant colony optimization, bat algorithm etc. The author will discuss all the above mentioned algorithms in detail with the help of supportive statements, case studies, problems and solutions.

170

S. R. Mani Sekhar et al.

9.3.1  R  elationship Between Traditional and Bio-Inspired Data Security The computing techniques which were used previously are good but still they require some changes for handling complex problems, they are not efficient at handling hard computations, pattern recognition, robust sensitive etc. Thus by using biological concepts with older techniques we can solve different kinds of complex problems and can enhance the robustness and efficiency too. Here we discuss bioinspired computing techniques in detail, its purpose and use, computational methods used applications and restrictions as well. Although existing traditional hardware and software can not completely fulfill all the needs and failed to solve many of the problems. The main disadvantage of traditional techniques is well stated but still some hard problems such as NP hard problems (TSP Problem), intelligent human machine interaction, natural language understanding, real world autonomous robots, management and understanding. Some of the other options available are BioComputing which involves simulation of biological mechanisms, Quantum computing and DNA based computing. The main foundation of bio inspired computing is natural sciences, different theory, conforming algorithms and artificial life.

9.3.2  Achieving Security in IoT Using Bioinspired Techniques The IoT is in its prime time involving different cities, smart homes, modular kitchens, smart buildings, smart appliances, smart health, smart machines and factories, smart transportation, and security. The management and buildings are based on smart systems consisting of sensors and participation of people as sources of information (Alam et al. 2013; Bello and Zeadally 2014; Perumal et al. 2016; Xu et al. 2014). The success of IoT in coming future depends upon the various new approaches in the field of evolution of technology and system security. The success of IoT is analyzed with the help of development in the field of system engineering and its processes along with the tools helping in overcoming the challenges in IoT network. The document discusses some bio inspired techniques for upcoming technology in communication in IoT platform which is an example of engineering in software. We talk about various optimization techniques such as computational iterative, which resulted in algorithms derived from nature, which are further derived from living creatures or organisms. Then further we explain examples obtained from the intellect of various existing organisms in nature. Bio inspired algorithm classifies evolution, ecology, immune, network, swarm in order to overcome difficulties such as security, scalability, mobility, heterogeneity, and resource constraint. The major goal of this chapter is to overcome and mention the emerging technologies for cyber security. This is used to prove that bio inspired techniques can really be used and further implemented to protect networks.

9  Bioinspired Techniques for Data Security in IoT

171

Biologically inspired models used in security have been flourishing for defending and securing ad hoc networks.

9.3.3  Types of Bioinspired Computing Algorithms Computing inspired from nature tells about the variety for different solutions which we can learn from our natural surrounding and thus implement it for solving problem related to computing. The world is full of different kinds of creatures and organisms thus by knowing about their behavior and characteristics one can learn and apply those concepts in computing for solving problems such as the NP hard problem, travelling salesman problem. The most popular algorithms inspired from nature are the ant colony algorithms, cuckoo algorithm, firefly algorithm, bee colony algorithm, bat algorithm, fish swarm algorithm and various others. These algorithms are developed by noting down the various characteristics of ants, bees, bats, and various other living creatures on which these algorithms are designed. Thus in this chapter we represent a few of these algorithms along with their case studies for better understanding and thus show how data security can be achieved in IoT using algorithms which are inspired by nature.

9.4  D  ifferent Approaches for Data Security in IoT Using Bioinspired Computing Bio-Inspired Computing involves the idea of natural computing which provides a field for performing research focusing on biology as an inspiration to solve complex computational problems along with using the experiences of the nature and natural world to solve real world problems. The various approaches used in bio inspired computing are described below. The major topics of bio inspired computing are as follows: (i) Ant colony Optimization (ii) Genetic Bee colony algorithm (iii) Firefly algorithm

9.4.1  A  nt Colony Optimization (Birattari et al. 2007; Dorigo and Birattari 2011) ACO consists of algorithms; the first one is called Ant System, which was given by Colorni, Maniezzo and Dorigo. The objective work, which is derived by the functions of real ants, has a similar find greater than various conforming gear which are

172

S. R. Mani Sekhar et al.

based on some restricted problem data which can be used with a active memory formation and has the information about the value of precious results. This kind of approach coming from the conversation of the diverse search gear has proved vital and helps in solving combinatorial optimization problems. In the past few years, the interest of the scientific community in ACO has grown immensely. The first of the ant algorithm, named “Ant System” (AS), was given by Dorigo et al. in 1996 and successfully tested on the popular problems. The ACO Meta or data definition was given by to generate, the method (Dorigo and Di Caro 1999) of conforming combinatorial problems by aggregate solutions and the solution to these problems is obtained from the amount of the bee interactions. A randomly spread first population result is spread in the dimensional problem space. A designated bee provides a change to the location in the reminiscence given to the local result and determine amount of nectar required for the new source. Then there are various phases involved in which the following part are involved Reproduction, in which based on the availability of the food the artificial bees waiting chooses a food source. After choosing the nectar amount the bees verify with the previous results if the new one is effective or needs to change to get optimized nectar value which is the required end product. In the other phases we check if the functioning of the above approach is what is required or not and check if other solutions are possible to solver the same problem. If the food source in the first phase is not been found, we move to other approaches in which the same concept is applied in a different way. In the next section we discuss the different case studies and present a broader perspective of what is being talked about here. Below section illustrates the different case studies such as Routing Protocols based on ACO, Ant Colony for travelling salesman problem, Bee Algorithm in data security, Dependable data assembly in IoT using bee colony, Performance analysis of firefly algorithm for data Clustering High Performance Security System for Image in IoT. 9.4.1.1  R  outing Protocols Derived from Ant Colony Optimization for Data Security in IoT (Liu 2017) Routing which is of course most important part of network based on wireless sensors and has been in the eyes of the research groups for quite a few years. Routing derived from ant colony is a kind of broadcast methodology which has various characteristics. There are various kinds of chapters that give and relate routing protocols from various views, but the assessment on ACO derived routing is still not there. This chapter makes an effort to provide a detailed review on these routing protocols which is based on ACO. First we give a detailed report on what these algorithms are and what are their functions and further classify them. Second we find the best suited ACO based protocols and describe them in detail for further analysis.

9  Bioinspired Techniques for Data Security in IoT

173

Types of ACO Inspired Routing in WSNs These algorithms are of various types. In this part, we give a comprehensive distribution of protocols which are mainly based on ant colony. The classification is given from different perspectives and is as follows. Behavior of Operation The ACO derived protocols are mainly of two types which are direction-finding of Energy Level Manage and Transmission Distance Manages. The above two types can be used depending on the type of behavior the operation has. Main Aim Based on the key objective, the routing protocol can also be separated into types which are Routing of life span addition and QoS requirements. The main objective of the first one is to increase the network duration, whereas the other aims at providing special QoS services. Topology of Network In the first type that is flat routing every node does the same work of path finding. Whereas in the other type that is Hierarchical routing many nodes are grouped to form clusters and path finding is carried out within these clusters. Probability Transition Based on the protocols of transition probability of ants, this type was derived and thus we say that it can be classified into two types Pheromone based and Heuristic based with only one pheromone. The first type the path finding is done by jointly pheromone trail and heuristic information. In this case study the author gave a detailed investigation based on any colony protocols. We have also made a distribution on the bases of these protocols. We have efficiently analyzed a good number of protocols based on ACO.

174

S. R. Mani Sekhar et al.

9.4.1.2  A  nt Colony Approach to Solve Travelling Salesman Problem (Beckers et al. 1992; Bolondi and Bondanza 1993) Here the chapter state ant colony capable of clearing famous travelling salesman problem. Ants of the artificial colony are made so that they make successively smaller tours which are feasible also by using the information which is gathered by the deposits on the sides of the TSP graph. The machine then give the ant colony is worthy of making good choices. These methods are an example much like the wireless networking protocols, network involving neural and competitive coding of the hopeful use of this symbol to design algorithms for the optimization. Ant Specific Algorithm In this type of algorithms artificial ant is like an agent which goes from one city to another on a TSP graph. It then decides on which city to move by using a probabilistic function which can used or given by both trail gathered on edges and of a heuristic value, which was selected to be like a function of the edges which determines the length. Artificial ants mostly prefer going to cities that are well attached by edges with a lot of pheromone trail and which are close by. Firstly, ‘n’ artificial ants are located on randomly picked cities. Each one at a given point in time step move to different cities and changes the pheromone trail on the edges which can use this to termed local trail from updating their locations. When all the ants have at least done one tour the ant that made the shortest tour has the right to change the edges belonging to its tour termed global trail changing by the addition of amount of pheromone trail that is inversely proportional to the length of the tour. The result obtained when these problems were tested and chosen mainly because there is a huge amount of data available to us in order to compare these results with historical. In the above case study, we compare the historic results which were either gathered from the bio inspired methods of the past and compare it with our new methodology. We find that our method is much more suited for solving NP hard problems and the famous travelling salesman problem as it is much more efficient to obtain the pheromone trail count or for that fact heuristic count to get the desired results. The key difficulty is to know the effective illustration for the problem and have an appropriate solution to it. In the experiments represented in this chapter local optimization was only used to enhance on the best results which we got from the various algorithms.

9.4.2  G  enetic Bee Colony (GBC) Algorithm (Alshamlan et al. 2015) Bee Colony Algorithm is an intelligence method (swarm) in which we select a gene hybrid method, which is GBC algorithm. The algorithm makes use of Genetic Algorithm in combination with use algorithms such as the ant colony algorithm. The

9  Bioinspired Techniques for Data Security in IoT

175

main aim which is focused is to combine the pros and benefits of both algorithms. A microarray gene expression profile is tested by applying this algorithm in order to obtain the gene which is the most predictive and gives more information for classifying cancer. For testing the accuracy and the performance of this algorithm, various experiments were also conducted. Three different binary data sets for a micro array were used, which consists of: leukaemia, lung and colon. Further, three different microarray datasets belonging to different classes were used, namely: leukaemia, SRBCT and lymphoma and then the results of the algorithm is also compared with the recently used technique. GBC algorithm was compared with various other algorithms which are closely related and that have been published recently with all the necessary datasets. Below section illustrate the different case studies related to Genetic bee colony algorithm. 9.4.2.1  B  ee Colony Algorithm Used for Data Security Using Routing System Protocols (Okdem et al. 2011) In case of a fire, a very quick decision is required as every second counts in a condition where a factor such as low visibility, high anxiety and low environmental familiarity exists. The models for fire evacuation routing that are existing further involves the use of various bio inspired algorithms like the ant colony algorithm or else the swarm algorithm, which can also analyze the latency and calculate the delay caused due to congestion during process of evacuation properly nor can it determine the best desired and best suitable layout consisting of signs indicating guidance to emergency exit; henceforth, bee algorithm is used and is thus expected to be able to solve this problem. It objective is to create and develop a model for fire evacuation for routing called the “Bee-Fire”, using bee colony algorithm for testing the model of routing. Bee-Fire is used for finding the optimal routing techniques for fire evacuation hence reducing the time of clearance and the net time for evacuation. Fire Evacuation Routing and Artificial Bee Colony Optimization (BCO) Fire evacuation is the process of movement from a dangerous place to a secure and a safer place. It is characterized into the following three phases: 1: validation of cue phase, 2: phase of decision-making, and 3: phase for refuge. The first phase and second phase are also called the pre evacuation phase whereas the refuge phase is called termed movement phase. The pre evacuation phase involves more decisions on survival than their actual movement. Decision should be made in a timely manner but in many situations evacuees get confused to follow a path. The main pattern in studying evacuation involves methods based on optimization that help creators to prepare evacuation plans with the help of various tools like scheduling model for optimization, routing models for optimization and network models for optimization. Evacuees use routes that they are familiar with. Table 9.1 discuss the evaluation of research in the field of human behavior in situation of a fire evacuation process.

176

S. R. Mani Sekhar et al.

Table 9.1  Evaluation of research in the field of human behavior in situation of a fire (Butler et al. 2017) Factors considered during analysis of evacuation of a building Methods used Spirit of the times Occupancy density and travelling speed are the two major factors. The latest factors including individual personality, per-son to person capacities and architectural design of buildings were also considered Velocity of movement Methods/ways used for documentation of the time required for purpose of movement were discussed The link to fire and human This factor told that size/area of fire is dependent on the behavior of behavior a personnel prior to or after commencing. The human behavior factor also came under discussion resulted that people pass through smoke to save other family members during escape. Evacuation of the Tall buildings were a key area of research during the phase of fire functionally impaired evacuation. Two methods are there: (1) proper use of lifts and elevators, (2) the strategy of defend-in-place. Studies regarding New and special factors were discussed here i.e. the flow rates in WTC/11 staircases, times for escape and pre-evacuation, pre-evacuation actions Time prior to evacuation It is found that time for pre evacuation is far away more significant as compared to movement to a safe and a secure place. It is found that a delay in escape and count of deaths are interrelated. Model for evacuation Evacuation simulation models were designed which later on emerged. A few of them were based on the behaviour of human during evacuation, focusing only on the ways and the time required to interpret data being documented Finding ways during Conduction of case studies is a very important way as an approach. evacuation Spatial connections, architectural designs, and layouts are marked, that they generate some unnecessary problems to the occupants.

As illustrate in Table 9.2, respondents who participated in the survey consisted of more females than males. A large portion of these respondents belong to age group between 19 and 30 years, followed by age group of 31–50 years. Both the groups that are below 18 and between 51 and 70 have around same number of respondents i.e. (10% or 9.7%). The educational levels were subdivided into 5 groups consisting of secondary/high school, pursue diploma, pursue a degree, pursue masters, and pursue doctorate. Most of these respondents 63.1% have completed their education(higher) with degree, followed by around 12% of them completing diploma, 10.7% doing masters, 8.7% doing secondary school or lower, and remaining 5.8% doing doctorate.

9  Bioinspired Techniques for Data Security in IoT Table 9.2 Demographic profiles of respondents (Olander et al. 2017)

Demographic characteristic Frequency Gender M 45 F 58 Age group Below 18 10 Between 19–30 66 Between 31–50 17 Between 51–70 10 Educational qualification Secondary/high 9 school Pursue diploma 12 Pursue degree 65

177

Percentage 43.7 56.3 9.7 64.1 16.5 9.7 8.7 11.7 63.1

Bee Colony Algorithm and Applications Bee colony in the form of swarm intelligence which are inspired by bee collective intelligence for honey bee collection was used to overcome the limitations in traditional techniques of optimization. BCO algorithm involves the functions of honey bees so that it can find nectar locations around their bee hives, where they communicate via waggle dance. BCO algorithm has an advantage over other algorithms that it involves multiple control parameters. Bees can be grouped as socially linked insects which stay together in their colonies in the form of a dynamical system. The two essential components in a bee system are food sources and for agers. The food source value depends mostly on factors like its concentration of energy, closeness to the hives, richness, and the ease of extraction, but its “profitability” is generally given by the amount. Foragers are also categorized into three different groups (a) employed, (b) unemployed foragers, and (c) experienced. The unemployed ones are responsible to look out for sources of food for exploiting. Scouter and onlooker’s bees are the two types of bees which are unemployed ones. In this algorithm, half of the bees belonging to the colony are used as either employed and the other half are onlooker bees. Finding Fire Evacuation Route Using BCO Algorithm BCO algorithm helps in solving complex problems involving optimization and each of the artificial bees are responsible for generating at least a solution for a given problem. The position of food source near the bee hives gives a solution, which is potential and the amount of nectar reveals something about the quality, exactness and fitness regarding the solution. This algorithm works by assuming that most of

178

S. R. Mani Sekhar et al.

the bees (around 50%) in the bee hives are of onlookers category and other 50% are of employed bees category. Also, it is assumed that the total population of onlooker bees and employed bees gives the total number of solutions. Initially, BCO algorithm creates a random distributed population solution which gives the position of food source. Each solution zi is a Dth -dimensional vector where D is number of products of input and cluster size for each and every data set. After initialization process, the positions population undergoes repeated cycles represented as one of the search processes formed by the three types of bees. The scout and employed bees choose their food sources randomly with same probability of each and every food source or the route of fire evacuation. Onlooker bees choose the food sources which are based on the probability which is proportional to the nectar amount in the food source. Optimal Routing Solution for Fire Evacuation This algorithm finds probability for every such possible solution for routing, which tells about the loyal nature of a bee which is employed towards a particular source of food. The solution having the least probability value is taken as the most close or the near fire evacuation route and an optimal one. The evacuees can escape from any emergency exits and need not select a path for the emergency exit having the smallest distance. The result was obtained through a model for fire evacuation routing “Bee-Fire” developed by applying some ABC algorithm for finding optimal fire evacuation routes for evacuees; hence reducing total time for evacuation. The model can further determine the best location for emergency exit guidance signs. One limitation is regarding that the software do not simulate the levels of emotions and evacuees reaction during fire hence requiring updating. Another disadvantage is regarding the volume exclusion effect which is not guaranteed according to the model. Also, the model is verified without the use of real human experiments and relying on enforcing people choice for routes which is very difficult to get in real life. It is advised to take into account factors such as the behavior and nervousness level of evacuees in case of a fire to make the simulation look more accurate. 9.4.2.2  D  ependable Data Gathering in IoT Using Bee Colony (Najjar-­Ghabel et al. 2018) The Internet of Things (IoT) technology helps and allows local users (devices) to converse with one other for sharing, preparation and meeting, hazardious warnings and some crucial information with no any human interference. In the case of emergency applications in IoT technology, an important aspect is to provide an suitable method for gathering data. The proposed solution in the approaches that are existing is based on construction of a spanning tree over the IoT devices and collecting data using the tree.

9  Bioinspired Techniques for Data Security in IoT

179

The disadvantages of these algorithms are that they do not take into account or consider the probability of device’s mobility or failure. These trees are arranged and are based on the desired preferences and are then later used for gathering data in succession. Each and every tree is a employed one and used for gathering data by splitting the preceeding one. The studies are divided into two different groups. The first group consists and involves algorithms regarding routing, among the devices, that have not used a tree structure. The approaches have considered energy efficiency and reliability as measurements. One more category is based on a scheme based on tree building approach, which acts as a backbone for data tree transmission to the base main station in systems of IOT. The Proposed Model The IoT principle is displayed as a graph of undirected nature denoted by G = (Vertex, Edge) where Vertex, represent the collection of nodes containing the devices used and the base main station, and Edge denotes links which are there between the nodes. Every node ni has two of the main properties: pi: Devices are moving in a random motion. The probability of mobility of the IOT devices can be found out based on their past records. The probability of mobility of ni, given by pi is a random number belonging to the range of [0,1]. This probability of base main station is set to null or 0 which tells that the mobility of the node is assumed to be 0. ei: The variable gives the idea of residual energy of ni. The energy value of the base main station is considered to be infinite. This algorithm is based on the construction of a spanning tree over the network. The spanning tree tk, has a feature namely mpk given by: mpk: the mobility probability of internal nodes. The algorithm uses the most resident devices that are selected as internal nodes for increasing the reliability of trees. Reliable Spanning Tree-Based Data Gathering in IoT It includes the approach consisting of tree construction which are reliable in IoT systems. The algorithm creates a set of spanning trees that are reliable to gather data. Firstly, data gathering is carried out on the most suitable tree and it is operational until the devices which are internal lacks energy. It leads to the process of tree splitting and further interruption of data gathering. Therefore, the next most suitable and reliable spanning tree will be used next. This continues till all trees have been utilized and employed. Here the author discussed the problem regarding gathering data that is reliable in IOT systems. Our algorithm that has been proposed, namely RST-IOT uses a struc-

180

S. R. Mani Sekhar et al.

ture of a tree for carrying out collection of data. In order to achieve the desired solutions with throughput being high, we modified the ABC algorithm leading to construction of a tree problem. The advantage regarding this is related to the fact that a swarm optimization algorithm which generates and results in many near-­ optimal trees. The amount of nectar in each and every solution is found using a few number of metrics such as residual energies of the devices, the number of hop distances count between various devices and base station, and the probabilities of their mobilities. A number of spanning trees are generated using the ABC algorithm. The trees are sorted based on the respective preferences and each of them is used to gather data. The results of simulation indicate that RST-IoT algorithm is superior to other approaches used in many terms such as reliability and consumption of energy. Hence it is a suitable way in handling emergency applications in IoT technology.

9.4.3  Firefly Algorithm (Yang 2008) The firefly algorithm (FA) is an algorithm inspired by nature which depends on the fireflies and their behavior of flashing. This algorithm is inspired by various features of fireflies. They produce a clipped and pulsating flash which attracts their mate partners thus serving as a mechanism or tool for protective warning. Firefly Algorithm utilizes and formulates this lightning behavior of lightning bug. The swarm of fireflies helps to solve the problem in an iterative way. The fitness value helps in giving information regarding the attractiveness of each member participating in the swarm. It is represented in terms of light intensity. The fireflies involved are firstly dislocated and further each and every firefly needs to find their mate depending on attractiveness of the rest of fireflies. They further approach their partner for improving condition and fitness. The performance is improved with the help of the F-Clust algorithm. A firefly is allowed to proceed towards its mate only if there exists a scope of improvement in the intensity value. 9.4.3.1  A  nalysis of Performance Using Firefly Algorithm Used for the Purpose of Data Clustering (Banati and Bajaj 2013) A very important technique which is being used these days for organizing information and data is called clustering. The problem of clustering refers to the process in which partitioning of data objects that are unlabeled, takes place into a number of clusters with aim to maximize homogeneity within the clusters as well as heterogeneity between clusters. The algorithms performance obtained is made to compare with PSO algorithm and DE algorithms in terms of statistical criteria which make use of various data sets. It is difficult to discover and extract relevant information due to the larger growth of data and information due to the popularization of web. The information extracted is based on the desired results of methods which are used to represent, organize and

9  Bioinspired Techniques for Data Security in IoT

181

access information at a same time. One such method used is clustering in which organizing of data takes place which facilitates the process of extraction of information (relevant). The problem of clustering involves the partitioning of data objects that are unlabeled into limited clusters. As the aim of clustering is focused at achieving homogeneity among clusters and then achieving heterogeneity among clusters, secondly and this task of the algorithm is defined as the optimization problem. FA is generally a nature based inspired (bio inspired) approach used for solving of optimization problem which are nonlinear and which has been introduced recently (Yang 2008). It is based upon the physical related behavior socially related insects (fireflies) where each of the fireflies have their own purpose, agenda and coordinate values with respect to the various similar or dissimilar fireflies in that group in achieving the same as shown in Fig. 9.1. A new and modified clustering approach for formulating the flashing/lightning behavior displayed by fireflies which are having the clustering problem’s objective function. The study has helped in calculating the performance of FClust algorithm in terms of accuracy and efficiency by carrying out comparison between the results of PSO and DE. Python language is used for the execution of algorithms and the other activities are carried out on a personal computer (PC). The effectiveness and its impact are calculated using following criteria: (i) The value of best mean fitness (ii) Success Percentage (iii) Hoose and Stutzl (2004) proposed run length distribution (RLD). Analysis of Performance These analyses are done for getting better performance of the used algorithms and can be generally be divided in to the following types. Using Artificial Data Sets The value of mean best fitness is found using artificial data sets and is carried out for 20 runs for analysis as compared with the artificial data sets are reported and recorded in tabular form and then compares the results with that of PSO and DE algorithms. Using Real World Data Sets The values of mean best fitness is carried out for 20 runs for the analysis as compared to the bench marked data sets are reported and recorded which are further calculated for each data sets such as that of irises.

182

Fig. 9.1  Firefly algorithm flowchart (Yang 2008)

S. R. Mani Sekhar et al.

9  Bioinspired Techniques for Data Security in IoT

183

Percentage of Success The rate of success in given as percentage of the number of runs, which gives the best value for the objective function carried out over 20 simulations. The result shows that the algorithm used in clustering called, FClust is responsible for dealing with the firefly algorithm and its behavior in order to improve and modify the existing solution to the given problem occurring in clustering. The performance of FClust is calculated by using the result obtained by the comparison between PSO and DE with the help of various datasets. The experimental study has proved that FClust algorithm is having a probability higher than algorithms such as PSO and DE in achieving optimality. 9.4.3.2  S  ecurity System for Image in IoT with High Performance (Alam et al. 2013; Suwetha et al. 2017) Exchanging image and video has become nowadays, one of the most important requirements to be used in the era of IOT. A Quad rotor is an application of IoT. In this type of rotor, images can be captured with Secure Digital Camera. After capturing of images, it is shared with the users who are intended, with a lot of security measures. It is achieved by using a technique of image hiding where an image which is secret is made hidden into another image (cover). After embedding and inserting the former image, the latter image will only appear. It is proposed for integrating and combining several algorithms with SDC for achieving security and long ranged transmission of images that are captured successfully. The algorithm used for integrating with SDC is called firefly which is used for hiding images. Thus an attempt is made to eliminate the concerns regarding privacy and security with the use of the integration of hybrid algorithms with SDC. Methodology Images that are captured in SDC have to be further transmitted with a lot of security for carrying out long ranged transmission; it is achieved by the integration of algorithms which are hybrid, with SDC like the firefly algorithm used for image hiding. The algorithm that has been proposed offers protection as a two layer consisting of encryption and hiding images, which is responsible for covering issues that are related to domains such as security purpose, privacy of users and management of digital rights (DRM). The novel contributions are: • • • • •

Quality of the image is improved During the process of image hiding, less memory and space is acquired Peak signal, which is high and a noise ratio (PSNR) is being achieved here. Reduced Noise is attained for hidden image. Mean Square Error (MSE) is acquired.

184

S. R. Mani Sekhar et al.

• Lossless compression is implemented. Henceforth, the SDC with the various hybrid algorithms is considered the most suitable and proved ways for facilitating the management of rights in the real time and is considered to be successful for applications in real time such as IoT. Image Hiding Using Firefly Algorithm The FA is basically an iterative technique in which the system which is proposed results or infers a new domain of frequency FA which depends on technique involving image hiding. The key idea involved in the technique is a two-fold idea: representation of image in multi resolution and odd-even quantization. The secret image is inserted into an another image called the cover image by the method of odd-even quantization for modifying the coefficients. Secret image is further encapsulated into the cover image with the help of a key image. Process of Hiding Images The two images, cover and secret images are categorized and divided into several 8 × 8 size non overlapping blocks. Each iteration involves finding an optimum location by the firefly algorithm and then the position of each firefly is updated. The most suitable location is found in the occurrence of certain conditions given below: • When the performed number of iterations exceeds the maximum desired iterations being carried out. • When there exists no any improvement obtained in the carried out iterations. • A desired, acceptable and suitable result is found. The secret image blocks are embedded in the most suitable possible pixel of every cover image block present and histogram shifting method is used for this purpose. The histogram denoted by H(x) is generated. The maximum intensity pixel is also known as the peak point while the pixel with minimum intensity is also called zero point which can also be found using a histogram denoted as H(x). Peak/highest point of image block which is secret is inserted in the peak point of the second image block called the cover block. In a similar way zero point of a secret image block is inserted into zero or the null point of the cover image block. After the completion of this process, SSIM (Structural Similarity Index Measure) is calculated for cover along with the embedded image. After this the BER value is analyzed and calculated. The SSIM describes the quality of embedded image. The Bit Error Rate (BER) gives information about how much robust an embedded image is. The BER value is between 0 and 1. If it is nearer to the value of 1 then the value of error for the extracted image is found to be higher. The sole purpose of the process is minimizing the BER and to maximize the SSIM ratio.

185

9  Bioinspired Techniques for Data Security in IoT

In this section, the author has developed hybrid algorithms which are implemented here with SDC accomplished in the work has successfully been done for carrying out image hiding, compression, long ranged communication and encryption efficiently. The algorithms have been classified and analyzed in reference to the parameters for evaluation of performance. The parameters used here, displays that the concerns in privacy as well as security in secured transmission of data could be obtained well as shown in Fig. 9.2.

Start

Initialize the number of fireflies

Find the value of fitness for each firefly

Fireflies are ranked based on their fitness values

YES If f(j)>f(i)

The ith firefly is moved towards the jth one

NO Find the fitness value for each firefly

Itr=itr+1

NO Stopping condition

YES End

Fig. 9.2  Flowchart displaying the image hiding process (Okdem et al. 2011)

186

S. R. Mani Sekhar et al.

9.5  Conclusion This chapter presented the work in the field of data security in IoT using bio inspires techniques, based on the behavior study of living organisms found in the nature. Subsequently it also illustrates algorithms used to solve real world problems in the case of data security in IoT. As mentioned in the chapter, the ant colony optimization algorithm is used to solve instances in data security in IoT such as routing protocols in wireless sensor devices which can be extended to various applications in this world through IoT. Also, ACO is used in TSP i.e. travelling salesman problem, which benefits the salesman in his sale. Similarly, artificial BCO algorithm is used in solving issues related to evacuation system in case of fire. As very less time is present to make a decision in the case of a fire occurs. To overcome this issue ABC algorithm is used in helping the evacuees to find the best optimal path for the emergency exit. In later case studies ABC algorithm is used in gathering data for the internet of things. Finally, firefly algorithm uses the flashing characteristics of the firefly in solving problems such as performance analysis for data clustering and a high performance security based system for an image. The above algorithms mentioned and discussed along with the case studies involved gives a supporting hand for the cause of achieving data security in Internet of Things with the help of biologically inspired computing techniques. The mentioned case studies thereby prove that bio-inspired computing can be used as a option of better solution for security in IoT.

References Alam, B., Doja, M. N., Alam, M., & Malhotra, S. (2013, September). Security issues analysis for cloud computing. International Journal of Computer Science and Information Security, 11(9), 117–125. Alshamlan, H. M., Badr, G. H., & Alohali, Y. A. (2015). Genetic Bee Colony (GBC) algorithm: A new gene selection method for microarray cancer classification. Computational Biology and Chemistry, 56, 49–60. Banati, H., & Bajaj, M. (2013). Performance analysis of firefly algorithm for data clustering. International Journal of Swarm Intelligence, 1(1), 19–35. Banu, R., Ahammed, G. A., & Fathima, N. (2016). A review on biologically inspired approaches to security for Internet of Things (IoT). In 2016 International Conference on Electrical, Electronics, and Optimization Techniques (ICEEOT). IEEE. Beckers, R., Deneubourg, J. L., & Goss, S. (1992). Trails and U-turns in the selection of the shortest path by the ant Lasius niger. Journal of Theoretical Biology, 159, 397–415. Bello, O., & Zeadally, S. (2014). Intelligent device-to-device communication in the Internet of Things. IEEE Systems Journal, 10(3), 1172–1182. Birattari, M., Pellegrini, P., & Dorigo, M. (2007). On the invariance of ant colony optimization. IEEE Transactions on Evolutionary Computation, 11(6), 732–742. https://doi.org/10.1109/ TEVC.2007.892762.

9  Bioinspired Techniques for Data Security in IoT

187

Bolondi, M., & Bondanza, M. (1993). Parallelizzazione di un algoritmo per la risoluzione del problema del commesso viaggiatore. Masters thesis, Politecnico di Milano, Italy. Croes, G.A. Butler, K., Kuligowski, E., Furman, S., & Peacock, R. (2017). Perspectives of occupants with mobility impairments on evacuation methods for use during fire emergencies. Fire Safety Journal, 91, 955–963. Dorigo, M., & Birattari, M. (2011). Ant colony optimization. In C. Sammut & G. I. Webb (Eds.), Encyclopedia of machine learning. Boston: Springer. Dorigo, M., & Di Caro, G. (1999). Ant colony optimization: A new meta-heuristic. IEEE. 2. 1477 Vol. 2. https://doi.org/10.1109/CEC.1999.782657. Liu, X. (2017). Routing protocols based on ant colony optimization in wireless sensor networks: A survey. IEEE Access, 5, 26303–26317. Najjar-Ghabel, S., Yousefi, S., & Farzinvash, L. (2018). Reliable data gathering in the Internet of Things using artificial bee colony. Turkish Journal of Electrical Engineering and Computer Sciences, 26(4), 1710–1723. Okdem, S., Karaboga, D., & Ozturk, C. (2011, June). An application of wireless sensor network routing based on artificial bee colony algorithm. In 2011 IEEE Congress of Evolutionary Computation (CEC) (pp. 326–330). IEEE. Olander, J., Ronchi, E., Lovreglio, R., & Nilsson, D. (2017). Dissuasive exit signage for building fire evacuation. Applied Ergonomics, 59, 84–93. Perumal, T., Sulaiman, M. N., Datta, S. K., & Leong, C. Y. (2016). Rule-based conflict resolution framework for internet of things device management in smart home environment. In IEEE global conference on consumer electronics. Sari, I. R. F. (2017, October). Bioinspired algorithms for Internet of Things network. In 2017 4th International Conference on Information Technology, Computer, and Electrical Engineering (ICITACEE). IEEE. Suwetha, K., Vinothini, K., & Shankar, R. (2017). High performance in security system for image in IoT. International Journal of Engineering Development and Research, 501–507. Xu, L. D., He, W., & Li, S. (2014). Internet of Things in industries: A survey. IEEE Transactions on Industrial Informatics. Yang, X. S. (2008). Nature-inspired metaheuristic algorithms. Beckington: Luniver Press.

Chapter 10

A Chaos-Based Multi-level Dynamic Framework for Image Encryption Sakshi Dhall, Saibal K. Pal, and Kapil Sharma

Abstract  Large scale internet and mobile usage has led the world witness ubiquitous increase in data transfer requirements as well as nature of data in transmissions on insecure networks. Internet of Things (IoT) has given an altogether new dimension to security challenges required to be addressed in the emerging world of smart cities specially with respect to data which is not necessarily text based and is magnanimous and requires to be processed in environments with adaptive needs. Images form one such type of data which forms significant proportions of modern day transmissions and is also equally significant when it comes to data being generated by devices including mobile phones, IoT devices like smart bells, surveillance devices, CCTVs etc. specifically when operating in environments requiring adaptability to dynamic changes in security requirements in balance with resource availabilities. Traditional schemes with proven security like AES use static operations involving significant computational expense and are suitable for securing textual data but no such standard exists till date for securing media like images. Since, images are bulkier and contains significant correlation in neighbourhood so there is a need to design new encryption approaches going beyond the conventional static encryption paradigm, hence, in this chapter we propose a paradigm shift opposed to the static approach for encryption. This chapter proposes a chaos-based, multiple-round, adaptive and dynamic framework for image encryption with new levels of dynamism across different functional dimensions of the entire encryption process. The impact of the new levels of dynamism across the entire encryption process has been experimentally demonstrated. Observations and results show that such framework can be used to address S. Dhall (*) Department of Mathematics, Jamia Millia Islamia, New Delhi, India Department of Computer Science & Engineering, Delhi Technological University, New Delhi, India S. K. Pal Scientific Analysis Group, DRDO, New Delhi, India K. Sharma Department of Information Technology, Delhi Technological University, New Delhi, India e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Alam et al. (eds.), Internet of Things (IoT), S.M.A.R.T. Environments, https://doi.org/10.1007/978-3-030-37468-6_10

189

190

S. Dhall et al.

dynamically changing encryption requirements and resist higher levels of cryptanalysis because the proposed dynamism of the framework makes it difficult for the attacker to identify and estimate the structure and operations of the encryption process to perform cryptanalysis. Keywords  Image encryption · Multimedia · Chaos · Dynamism · Adaptive · Dynamic framework

10.1  Introduction Improvement in price-performance ratio of microelectronics technology has given a strong impetus to what we call as digital revolution. Digital information has found its inevitable presence in almost every sphere of life from military to medical science, banking to e-commerce, education to entertainment, tourism to social media to space exploration etc., and the list is endless. The digital technology has no doubt proved to be a marvel in its contribution to the kind of communications possible in today’s world but it has also paved way for serious possible breaches when sensitive information is stored or shared through susceptible channels. With the offshoot in mobile technology and internet becoming more and more permeant part of our everyday lives, the magnitude of available digital data and the need for voluminous information transfers have increased enormously. Applications like financial transactions, VoIP, medical and military imaging, social networking, real-time streaming, pay-TV, video conferencing etc. have resulted in bulk data exchanges to become indispensible for our professional and personal activities. These data exchanges may include copyrighted and confidential information exchanges including multimedia esp. images (Furht 2005; Gonzalez and Woods 2007) making it highly important to have impenetrable channels and safe end points so that digital data transfer and storage happens in a secured way. As security of digital content impacts diverse sets of people ranging from large business houses to defense establishments to space research organizations to health care sector to entertainment industry to individuals and so on, it is imperative to equip all information generating and transmitting systems with appropriate data securing capabilities. This will ensure that the communicating parties feel safe and secure while exchanging confidential information over otherwise penetrable networks thereby providing them an ease and confidence in carrying out the required communication. Cryptographic techniques (Stallings 2004; Menezes 1996) like encryption play a vital role in securing confidentiality of sensitive information by converting them into non-perceivable forms thereby ensuring that any attempt of unauthorized access does not disclose any meaningful information to the adversary. Encryption refers to the process in which the meaningful plaintext is converted to the cipher text (non-perceivable form) using key and the reverse process of obtain-

10  A Chaos-Based Multi-level Dynamic Framework for Image Encryption Fig. 10.1 (a) Basic encryption process. (b) Basic decryption process

191

(a) Plaintext

ENCRYPTION

Cipher text

Key (Secret)

(b) Cipher text

DECRYPTION

Plaintext

Key (Secret)

ing back the plaintext from the cipher text is referred as decryption. Figure  10.1 shows the basic encryption and decryption processes in symmetric-key cryptography. As evident, the digital data being mentioned is not restricted to textual content anymore. It contains significant amount of multimedia of which visual content forms a significant core. In fact, multimedia like images, video etc. has emerged out as major sources of digital content prevalent in storage and on networks today. With the increasing multimedia applications trends, the area of multimedia security esp. image security is also evolving and its scope has widened. Some of the applications requiring security measures for different forms of visual information are: • Images – forensics, defence maps, biometrics etc. • Video – surveillance, live-transmissions (esp. for defence, embassies, policing etc.), video-conferencing, DTH, social media etc. Further, with the frequent advancement in the area of Internet of Things (IoT) and machine to machine communication, the requirement of securing all forms of data including visual content like images have increased manifolds (Rohokale and Prasad 2015; Meskanen et al. 2015; Abomhara and Køien 2015; Elmangoush and Magedanz 2017). With the concept of smart cities becoming reality, intelligent IoT devices are required to securely communicate visual content like images with each other for various applications like surveillance, medical imaging, expert systems etc. At the first look, it appears that the conventional techniques, designed for securing text, should in principle be directly applicable on any form of data but, this does not hold true in practice. The reason is that multimedia data like images, videos etc. have special characteristics, i.e. they are bulky in size and possess strong correlation among neighbouring data bytes. These characteristics were missing in textual data, hence, these characteristics do not get directly addressed in traditional encryption schemes (Stallings 2004) like AES, DES, IDEA etc. when used in native ECB mode. Figure 10.2 show the original standard gray-scale ‘Water Lilies’ image and the corresponding cipher image obtained by encrypting it using AES in ECB mode.

192

S. Dhall et al.

(a)

(b)

Entropy: 7.9970

Textured Zones

Fig. 10.2 (a) Original image (water lilies). (b) AES encrypted image (water lilies)

This clearly shows that the cipher image contains textured zones because the ­redundancy in the input plain image does not get completely removed after performing the said encryption. These textured zones acts as vulnerabilities and can be easily be exploited by the attackers. Also, these traditional ciphers are static in their operations and secrecy is provided only by the private key. They largely rely on the mathematical complexity of the applied computations and secrecy of the key. They are computationally expensive and their static behaviour makes them susceptible to get broken, if not truly practically at least theoretically, when undergone significant analysis by the attacker (Phan 2004; Biryukov and Khovratovich 2009; Biryukov et al. 2009). Attacks are also proposed on AES family of cryptosystems even when used in, the otherwise considered secure, CTR mode (Biryukov et al. 2010). Also, in today’s time, many applications require securing multimedia content like images in resource constrained environments like mobile phones (Paul and Irvine 2016; Herland et  al. 2016; Ofori et al. 2016) and in such applications traditional schemes like AES will prove to be computationally more expensive. Also, when it comes to communications of visual content including those by IoT devices, some applications have high security requirements while for others efficiency may be a more significant factor and hence, practically such devices are operating in environments with varying/dynamically changing security requirements as per the application needs and resource constraints at hand. Therefore, there is a need for encryption schemes which are adaptable based on the changes in the operating environment and changing security needs. Also, this idea of adaptability of encryption scheme is well aligned and will be well utilized with the dynamic AI capabilities supported in the upcoming 5G technology for modern day communications.

10  A Chaos-Based Multi-level Dynamic Framework for Image Encryption

193

Though, there are several encryption schemes which have been proposed for securing multimedia including images yet there has been no encryption standard till date designed to cater to the special needs of multimedia content. In the light of traditional static standard schemes not being directly able to address security concerns for multimedia like images and absence of any image encryption standard, we feel strongly motivated to explore dynamic encryption as a paradigm shift for encrypting images. The approach can definitely be extended later to cater to other forms of multimedia as well. Introducing dynamism as part of the design of the encryption scheme does not take away the deterministic and reversible nature that every cryptosystem must possess. Despite offering dynamism in structure of the encryption process based on the key and also on the plaintext, the proposed work very well follows the Kerckhoff’s principle according to which every cryptosystem should be secure even if everything else about it is publically known except the secret key. Also, introduction of chaos-based dynamism can ensure that without using intricate mathematical operations, the dynamic framework could provide high security itself. Statistical and cryptanalytic attacks will intuitively become difficult when the encryption algorithm is dynamically changing with change in key and/or plaintext. This argument can be made stronger by highlighting that cryptanalytic attacks are normally designed by identifying some weak operation or loophole within the design of the static encryption scheme and then it is tried to be cracked by exploiting the identified loophole to extract partial or complete information about the plaintext or key or both (Özkaynaka et al. 2012; Tu et al. 2013; Wang et al. 2013; Norouzi and Mirzakuchaki 2016). But identifying such loopholes in a dynamic cryptosystem will surely be much difficult as the adversary will not even be equipped with the complete knowledge of the structure of the cryptosystem itself, which, of course, will be dynamically changing. Thus, chosen/known plaintext attacks will not be possible on such encryption scheme involving dynamism. In the following background section, a brief survey on existing subtle use of dynamism in encryption is also given. It shows that researchers have been suggesting use of dynamism in one form or the other as a means to enhance security of their proposed encryption schemes and to make it difficult for adversaries to attack. Extending the approach of dynamism, in this chapter, we propose a new chaos-­ based, multi-level dynamic framework for encryption using extendable 128-bit key and multiple rounds to encrypt images byte-by-byte. New dimensions to dynamism have also been proposed in this work, which are not employed in the existing works. It is proposed that the number of operations performed per round of the encryption scheme be decided dynamically. Also, the sequence of encrypting pixels will also be dynamically decided. This is incorporated along with dynamic decision on the type of operations performed. This kind of multi-level dynamism has not been used in existing literature and it will ensure that structure of each round of operations dynamically changes in an unpredictable way. This is done with an objective to increase the level of difficulty manifolds for an attacker to perform cryptanalysis to break the cryptosystem. In absence of knowledge of the key, the attacker will be completely unaware about the number and

194

S. Dhall et al.

nature of operations performed in the different rounds thereby averting cryptanalytic attacks. Also, as the type and number of per-round operations performed on the same pixel position will vary if the same key is applied with a slightly modified image hence proposed work will offer resistance against known/chosen plaintext attacks, differential and impossible differential attacks. Further, chaos has been made integral part of the framework to ensure high level of randomness in dynamism with less computational expense. Chaos provides properties like high sensitivity to changes in initial conditions and control parameters, ergodicity, random-like behaviour etc. (elaborated in the next section) which are very well suited to meet the confusion and diffusion requirements (Shannon 1949) of any cryptosystem. Hence, chaos has been widely used in cryptographic applications for decades now. Though, some researchers also criticize use of chaos as it requires floating point number processing but this criticism does not stand much, as today’s machines are well equipped to process real numbers with ease and efficiency. In fact, this very characteristic of chaos, i.e. playing with real numbers, actually possess great potential and makes it a strong contender to provide security solutions esp. for multimedia in post-quantum computers (Kartalopoulos 2010; Stojanovic et al. 2016; Geetha 2012; Akhshani 2015; Behnia et al. 2012; Akhshani et al. 2013; Ramos and Souza 2001; Ramos 2017). Further, it is to be highlighted that the proposed work is different from most of the other general chaos-based image encryption schemes due to its basic focus on use of chaos as a tool to generate dynamism in the encryption process, and not only as means to generate random sequences to be utilized in static permutation and substitution phases. The organization of the rest of the chapter is as follows. The use of chaos and dynamism in encryption as a background study is presented in Sect. 10.2. Further, Sect. 10.3 details the proposed dynamic encryption framework for image security. In Sect. 10.4, security analysis and the obtained results are elaborated. Section 10.5 envisions the future scope and concludes the chapter.

10.2  Background 10.2.1  Chaos in Cryptography Chaos (Devaney 1989) is a mathematical concept which offers special properties which can be well utilized in cryptosystems. Chaotic maps have non-linear dynamics, and they show random-like behaviour. They possess and show inherent avalanche properties as chaotic maps are highly sensitive to initial conditions and control parameters. Further, chaotic maps have high mixing property and ergodicity property which ensure that two initially very close points may get diverged after few iterations to points which are highly uncorrelated and the chaotic system uniformly

10  A Chaos-Based Multi-level Dynamic Framework for Image Encryption

195

attains different possible states without quickly converging to fixed value(s). These properties of high sensitivity, ergodicity, randomness allow use of simpler and computationally efficient operations to design cryptosystems with chaos. Chaotic maps may be continuous or discrete. The continuous maps can be generalized and discretized for application in encryption techniques. They can be used to generate random-like sequences which are normally utilized in key-stream generation, permutation and substitution steps of the encryption process. Due to ergodicity and the inherent avalanche property, the sequence generated by a chaotic function appears as a random noise with very small change in the initial conditions and/or control parameters. This adds to the unpredictability to the adversaries in absence of knowledge of the exact initial conditions and parameters. Further, as chaotic functions display deterministic behaviour therefore smooth recovery of the plaintext from cipher text during decryption is ensured. Initially, 1-D chaotic maps were only used by researchers but vulnerabilities were identified. So, higher order chaotic maps like 2-D, 3-D maps have become popular choice of researchers, to ensure better security without compromising on simplicity of operations. Some of the well known chaotic maps include: 1D: logistic map, tent map, sine map. 2D: Baker’s map, Arnold’s cat map, Hénon map, Ikeda map. 3D: Lorenz System, Chen-Lee system. Figure 10.3 displays sensitivity of a chaotic map namely logistic map with respect to the value of initial condition and parameter to the extent that highly uncorrelated orbits are generated by using the logistic map with same control parameter value i.e. set to 4 and minor change in the value of initial conditions i.e. x0 = 0.7 and x0 = 0.701. Both these plots represent the dynamics of the logistic map in terms of plots of 100 points obtained after skipping first 900 iterations to eliminate the transient effect  (http://s3.amazonaws.com/complexityexplorer/ NonlinearDynamics/Nonlinear.LogisticLabII.pdf). In fact, researchers have also been proposing new, improved or hybrid chaotic maps (Zhou et al. 2014; Ramadan et al. 2016; Boriga et al. 2014; Zhang and Cao x0 = 0.701

x0 = 0.7

Fig. 10.3  Plots for Logistic Map with minor change in the value of initial conditions (control parameter value as 4)

196

S. Dhall et al.

2014; Elabady et al. 2014; Saraereh et al. 2013; Borujeni and Ehsani 2015; Maqableh 2015; Rui 2015) by making improvisations or by combining more than one chaotic map to enhance the chaotic behaviour of existing maps, like, making the new proposed maps which possess longer orbits, wider range of acceptable parameter values showing chaotic behaviour, higher randomness etc.

10.2.2  Survey on Dynamism in Encryption Encryption process in any block cipher normally involves key-stream based permutation and substitution operations which are repeated over a number of rounds. There are various levels at which dynamism can be introduced in this encryption process like dynamism in encoding/decoding, dynamism in the number and type of operations used, dynamism in number of rounds employed, dynamically changing keys, dynamism during key-stream generation from the original key etc. Figure 10.4 pictorially represents these different components of the encryption process in a block cipher where dynamism can be introduced. The roots of dynamism can be observed from polyalphabetic ciphers proposed as early as sixteenth century like Vigenere cipher where the idea to use of different transforms based on the secret key first surfaced. In 1984, S.  Goldwasser and S.  Micali (1984) proposed a new approach for encryption called Probabilistic Encryption. The authors highlighted that the traditional ciphers are deterministic trapdoor functions using secret information as trapdoor for security. And there exists a finite possibility that full or partial information about the plaintext may get revealed when attacked by adversary. To address this, the authors propose a proba-

Plaintext

Dynamically changing keys

Dynamism in key-stream generation

Dynamism in Encoding

Dynamic Selection of number and type of operations

Dynamism in number of rounds

Dynamism in Decoding

Cipher text

Fig. 10.4  Components of encryption process in any block cipher where dynamism can be introduced

10  A Chaos-Based Multi-level Dynamic Framework for Image Encryption

197

bilistic framework replacing the static one. As per probabilistic encryption the same plaintext bit 0 or 1 could be encrypted in several encodings in the cipher text where a possible encoding is chosen through a random mechanism. It is evident that probabilistic encryption framework encourages dynamism in selecting possible encoding for the same message bits in the cipher text to make it tough for the adversary to attack. In (Faraoun 2010) a stream-cipher structure is proposed and use of dynamism is proposed for generating chaos-based pseudo-random key stream being used to perform encryption. Pareek et  al. (2005) proposed dynamism in deciding length of blocks, choosing one chaotic map (for encryption of the block) among four chaotic maps along with their initial conditions and the number of iterations made to the chosen chaotic map. Later, Wei et al. (2007) cryptanalyzed this work but did not criticize the dynamic approach, instead, they suggested improvements to strengthen the dynamism by making it plaintext-dependent besides being key-dependent. Such dynamism in choosing among multiple chaotic maps has been proposed in (Abuhaiba et  al. 2012; Wang and Qing 2009; Sakthidasan and Santhosh Krishna 2011; Zhou et al. 2013; Vahidi and Gorji 2014) as well. The papers (Ngo et  al. 2010; Harmouch and Kouch 2015) proposed use of dynamically changing keys for encrypting different data packets of a communication so as to provide strength against cryptanalytic attacks by adversaries and to ensure that a compromised key do not reveal much information about the communication thereby not causing much harm. Several researchers (Cao 2013; Wang et al. 2015; Murillo-Escobar et al. 2015; Armand Eyebe Fouda et al. 2014a, b; Wang et al. 2016; Chai et al. 2017; Khan et al. 2017a, b; Ahmad et al. 2017) have taken the idea of dynamism in keys to another level by making keys used in the image encryption algorithm dependent on plaintext or using information extracted from the plaintext (like average intensity value, hash value etc.) and utilizing it in the image encryption operations thereby using this plaintext related information indirectly as the key which definitely needs to be communicated to the receiver along with the key for smooth decryption at the receiver’s end. Communication of such plaintext related information in a secure manner to the receiver is required for smooth decryption. It can be done either by hiding such plaintext related information inside the cipher text itself (Cao 2013) or otherwise securely communicating it with the key, which is not always practically feasible especially in resource constrained environments. Pareek et al. (2006) proposed a block cipher suitable for real time image encryption. The block cipher involves dynamism by making chaos-based selection of operations to be performed for encryption of each pixel of the image. In another work, Dhall and Pal (2010) proposed a 128-bit block cipher based on key-based conditional encryption. The algorithm comprises of two steps i.e. re-adjustment phase and substitution & shifting phase whose operations vary based on the key. This key-based conditional approach is then extended and adapted to make it suitable for multimedia in (Dhall et al. 2014). To suit multimedia, the block size is made customizable and three variants are suggested to ensure complete removal of redundancy in the cipher image. Later, Korstanje and Keliher (2015) cryptanalyzed the

198

S. Dhall et al.

scheme (Dhall and Pal 2010) by proposing distinguishing attacks and ­plaintext-­recovery attacks for keys with certain specific characteristics but did not criticize the conditional nature of choice of operation as such. Several encryption schemes like Blowfish (Schneier 1994), Twofish (Schneier et al. 1999) use key-dependent S-Boxes in the substitution step of the scheme. Key-­ based dynamic selection of S-Box for substitution is proposed in (Hussain et  al. 2012; Anees et al. 2014). Further, papers (Khan et al. 2017b, c; Ahmad and Hwang 2015) proposed improvisations on the scheme proposed in (Anees et  al. 2014). Also, several researchers in their works (Krishnamurthy and Ramaswamy 2008; Abd-ElGhafar et  al. 2009; Subramanyan et  al. 2011; El-Sheikh and El-Mohsen 2012; Juremi et  al. 2012; Mahmoud et  al. 2013; Pradhan and Bisoi 2013; Arrag et  al. 2013; Dara and Manochehr 2013) have proposed use of key-dependent S-Boxes in the standard algorithms like AES and have demonstrated and proved the strength of introducing key-based dynamism in this form. The idea of Dynamic Encryption has been discussed most explicitly by Knudsen in (Knudsen 2015) where it has been proposed that the while the performing encryption and decryption respectively at the sender’s and receiver’s ends, the receiver is kept unaware about the cryptosystem being used for encryption and only key is known to the receiver. The choice of cryptosystem being used for encryption is being kept to the sender who can change it as often as per message. To ensure smooth decryption at the receiver’s end, the sender communicates the executable code for decryption or encrypted decryption algorithm to the receiver along with the cipher text. On receiving the decryption algorithm along with the cipher text, the receiver can decrypt the message using the known shared key. Knudsen proposes practical advantages of such dynamic encryption for email systems, cloud storage and mobile conversations. This idea has been extended in our present work but with a difference. Unlike dynamic encryption proposed by Knudsen (2015), in our work the dynamic framework itself takes care of offering key and plaintext dependent dynamically changing encryption operations for pixels while the receiver is fully aware of the decryption algorithm i.e. there is no need to communicate or send the decryption algorithm separately to the receiver. Following section 3 describes our proposal. The use of dynamism in choosing rule/method of encryption has also been shown in some more recent works including (Husni 2017; Gmira et al. 2019).

10.3  Proposed Approach 10.3.1  Description of the Proposed Framework In this work, chaos has been primarily used to introduce random-like dynamism in the overall structure of the proposed framework. Each pixel is visualized to be composed of three parameters i.e. (x, y, color byte) and, as part of the encryption process, a new color byte value is to be generated for each pixel. To ensure complete

10  A Chaos-Based Multi-level Dynamic Framework for Image Encryption

199

removal of redundancy existing in the plain image, besides these three parameters, each pixel is assigned a key-based ‘chaotic value’, where 0